Estimating tumor purity from single samples

Abstract

To provide a method for estimating tumor purity from a tumor sample without usage of matched-normal control.SOLUTION: A method is configured to: identify a set of genomic regions on the basis of nucleic acid sequence data that is aligned to a reference genome, in which each genomic region of the set of genomic regions includes one or more nucleotide-sequence variants related to a corresponding genomic region of the reference genome,; measure a B-allele frequency distribution of a biological sample on the basis of a B-allele frequency measured with respect to each genomic region of the set of genomic regions; process the B-allele frequency distribution using a trained machine-learning model; and estimate a metric identifying tumor purity in the biological sample.SELECTED DRAWING: None

Description

[Technical Field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims priority to U.S. Provisional Patent Application No. 62 / 931,096, filed November 5, 2019, which is incorporated herein by reference in its entirety for all purposes.

[0002] The present disclosure relates generally to systems and methods for estimating tumor purity from a single sample. More particularly, but not by way of limitation, the present disclosure relates to estimating tumor purity of a biological sample by processing B allele frequency distributions using a trained machine learning model. [Background technology]

[0003] Tumor cellularity, also referred to as "tumor purity," determines the proportion of cancer cells in a sample. Accurate estimation of tumor purity in a biological sample can contribute to improved accuracy in detecting somatic mutation abundance and / or copy number changes. This is because tumor purity indicates the allele frequency of somatic mutations present in a biological sample. Detection of somatic mutations and copy number variations can then be used to determine the stage of a subject's cancer or evaluate whether a particular cancer treatment is effective. Therefore, tumor purity can be useful in determining the stage of cancer and / or evaluating the efficacy of treatment.

[0004] While tumor purity can be an effective metric, it can also be a confounding variable in some bioinformatics analyses. For example, traditional approaches to estimating tumor purity may require histopathological evaluation by a pathologist by manually inspecting sample images to estimate tumor purity. However, histopathological evaluation, including manual inspection of sample images, can be subjective and inaccurate. Other traditional techniques for estimating tumor purity require comparing values (e.g., putative somatic mutations) derived from nucleic acid sequence data of a given tumor sample with other values derived from nucleic acid sequence data of a matched normal control sample. However, such normal control samples may not be available.

[0005] For example, conventional approaches estimate the tumor purity of a sample as a function of the allele fraction of somatic mutations unique to an individual's tumor. In the absence of a matched normal sample, identification of these somatic mutations becomes less accurate, significantly reducing the accuracy of the estimated purity. In some cases, a matched normal control is unavailable if (for example) the sample source did not collect or sequence a normal control.

[0006] Therefore, there is a need to accurately estimate tumor purity in a sample to facilitate detection without relying on subjective analysis (e.g., histopathological evaluation) or the presence of normal control samples. Summary of the Invention [Problem to be solved by the invention]

[0007] In some embodiments, a method for estimating tumor purity is provided. The method may include obtaining nucleic acid sequence data representing a plurality of nucleic acid molecules of a tumor sample from a subject. The method may also include aligning the nucleic acid sequence data to a reference genome. The method may also include identifying a set of genomic regions based on the aligned nucleic acid sequence data. In some cases, each genomic region of the set of genomic regions comprises one or more nucleotide sequence variants relative to a corresponding genomic region of the reference genome.

[0008] The method may also include measuring a B allele frequency for each genomic region of the set of genomic regions. The method may also include measuring a B allele frequency distribution for the biological sample based on the B allele frequencies for the set of genomic regions. The method may also include processing the B allele frequency distribution using a trained machine learning model to estimate a metric that identifies tumor purity for the biological sample. The method may also include outputting the metric.

[0009] In some embodiments, a system is provided that includes one or more data processors and a non-transitory computer-readable storage medium that includes instructions that, when executed on the one or more data processors, cause the one or more data processors to perform some or all of one or more methods disclosed herein.

[0010] In some embodiments, a computer program product is provided that is tangibly embodied in a non-transitory machine-readable storage medium and includes instructions configured to cause one or more data processors to perform some or all of one or more of the methods disclosed herein.

[0011] Some embodiments of the present disclosure include a system including one or more data processors. In some embodiments, the system includes a non-transitory computer-readable storage medium including instructions that, when executed on the one or more data processors, cause the one or more data processors to perform some or all of one or more methods and / or some or all of one or more processes disclosed herein. Some embodiments of the present disclosure include a computer program product tangibly embodied in a non-transitory machine-readable storage medium and including instructions configured to cause one or more data processors to perform some or all of one or more methods and / or some or all of one or more processes disclosed herein.

[0012] The terms and expressions employed are used as terms of description and not of limitation, and there is no intention to use such terms or expressions to exclude any equivalents of the features shown and described or portions thereof, but it is recognized that various modifications are possible within the scope of the invention as claimed. Thus, while the invention as claimed has been specifically disclosed by embodiments and desired features, it will be understood that modifications and variations of the concepts disclosed herein may be practiced by those skilled in the art. Such modifications and variations are deemed to be within the scope of the invention as defined by the appended claims. [Brief explanation of the drawings]

[0013] The features, embodiments, and advantages of the present disclosure will be better understood upon reading the following detailed description in conjunction with the following figures.

[0014] [Figure 1] FIG. 1 shows a schematic diagram of a fully connected neural network for estimating tumor purity according to some embodiments. [Figure 2] FIG. 2 shows a schematic diagram of a one-dimensional convolutional neural network for estimating tumor purity according to some embodiments. [Figure 3] FIG. 3 shows a schematic diagram of a two-dimensional convolutional neural network for estimating tumor purity according to some embodiments. [Figure 4A] FIG. 4A shows loss versus epoch plots for a fully connected neural network, a one-dimensional convolutional neural network, and a two-dimensional convolutional neural network trained on a VCF-level dataset, according to some embodiments. [Figure 4B] FIG. 4B shows root mean square error (RMSE) and mean absolute error (MAE) data for a fully connected neural network, a one-dimensional convolutional neural network, and a two-dimensional convolutional neural network trained on the VCF-level dataset, according to some embodiments. [Figure 4C]FIG. 4C shows root mean square error (RMSE) and mean absolute error (MAE) data for testing of the FASTQ level dataset, according to some embodiments. [Figure 5A] FIG. 5A shows a plot 500A comparing true tumor purity with tumor purity predicted by a trained fully connected neural network, according to some embodiments. [Figure 5B] FIG. 5B shows a plot 500B comparing true tumor purity with tumor purity predicted by a trained two-dimensional convolutional neural network, according to some embodiments. [Figure 5C] FIG. 5C shows a plot 500C comparing true tumor purity with tumor purity predicted by a one-dimensional convolutional neural network, according to some embodiments. [Figure 6] FIG. 6 shows an example of a schematic for estimating tumor purity using a trained machine learning model, according to some embodiments. [Figure 7A] 7A-7E provide examples of plotted BAF distributions according to some embodiments. [Figure 7B] 7A-7E provide examples of plotted BAF distributions according to some embodiments. [Figure 7C] 7A-7E provide examples of plotted BAF distributions according to some embodiments. [Figure 7D] 7A-7E provide examples of plotted BAF distributions according to some embodiments. [Figure 7E] 7A-7E provide examples of plotted BAF distributions according to some embodiments. [Figure 8] FIG. 8 shows a plot 800 of the first two major components for classifying biological samples using B allele frequency features, according to some embodiments. [Figure 9] FIG. 9 includes a flowchart illustrating an example of a method for estimating tumor purity of a biological sample according to certain embodiments. [Figure 10]FIG. 10 illustrates an example computer system for implementing some of the embodiments disclosed herein. DETAILED DESCRIPTION OF THE INVENTION

[0015] I. Overview To address at least the above-mentioned deficiencies of conventional systems, the techniques of the present invention can be used to estimate tumor purity by processing the B allele frequency distribution of heterozygous sites in sequencing data using a trained machine learning model. The trained machine learning model can generate an estimated metric to identify the tumor purity of a biological sample even in the absence of matching sequencing data from a normal control sample. Thus, the techniques of the present invention can accurately estimate tumor purity in biological samples without a matched normal control, thereby facilitating more accurate analysis of nucleic acid sequencing data from tumor-only samples.

[0016] Nucleic acid sequence data representing a plurality of nucleic acid molecules of a tumor sample of a subject can be obtained. In some embodiments, the nucleic acid sequence data is whole-exome sequencing data. The nucleic acid sequence data may be whole-genome sequencing data. In some embodiments, the sequencing data is from a tumor sample. The tumor sample may be from a human subject. The nucleic acid sequence data may be generated by shotgun sequencing. In some embodiments, the nucleic acid sequence data may be generated by sequencing a selected portion of the genome or exome.

[0017] The nucleic acid sequence data may be aligned to a reference genome. A set of genomic regions may be identified based on the aligned nucleic acid sequence data. In some cases, each genomic region of the set of genomic regions contains one or more nucleotide sequence variants relative to a corresponding genomic region of the reference genome. To identify the nucleotide sequence variants, candidate variants may be identified and a reference read depth and an alternative read depth for the candidate variants may be calculated.

[0018] The B allele frequency of each genomic region of the set of genomic regions can be measured. B-Allele Frequency (BAF) refers to a normalized measure of the allele intensity ratio of two alleles (A and B), where a BAF of 1 or 0 indicates the complete absence of one of the two alleles (e.g., AA or BB), and a BAF of 0.5 indicates the equal presence of both alleles (e.g., AB). Based on the B allele frequencies of the set of genomic regions, the B allele frequency distribution of the biological sample can be measured. In some embodiments, the B allele frequency distribution is normalized.

[0019] The trained machine learning model may be used to process the B allele frequency distribution to estimate a metric that identifies the tumor purity of the biological sample. As used herein, tumor purity or tumor cell content refers to the proportion of cancer cells in a tumor sample. Tumor purity can be a metric used in various techniques to estimate tumor characteristics that correspond to direct clinical relevance. For example, the estimated tumor purity may directly modify the threshold of evidence required to identify copy number alterations, including those associated with approved companion diagnostics (e.g., EGFR exon 19 deletion in NSCLC). Tumor purity estimates can also be used to improve the quality (sensitivity, PPV) of somatic variant calls by prioritizing the expected allele fraction of somatic variants. For example, a tumor purity value of 60% may indicate that a tumor sample contains 60% tumor cells and 40% normal cells. Additionally or alternatively, tumor purity may indicate the number of cancer cells in a tumor sample. Note that whenever the term "tumor purity" is mentioned, the term "tumor cell content" may be used interchangeably throughout this specification.

[0020] The trained machine learning model may be trained on a dataset generated in silico (e.g., nucleic acid sequence data generated by computer simulation). In some embodiments, the trained machine learning model is trained on a training dataset generated from sequencing data derived from tumor cells diluted with normal cells. The mean absolute error of the trained machine learning model may be less than about 0.2. In some embodiments, the root mean square error of the trained machine learning model is less than about 0.2.

[0021] In some embodiments, the trained machine learning model comprises a fully connected neural network. The fully connected network may include a fully connected layer with a rectified linear unit (ReLU) activation function. In some embodiments, the output activation function of the fully connected neural network is a sigmoid function. A loss function of the fully connected neural network may be configured to calculate mean squared error (MSE). In some embodiments, the fully connected neural network is tuned by hyperparameter search using random sampling with a linear search of layers, a linear search of size, a logarithmic search of learning rate, or a combination thereof.

[0022] In some cases, the trained machine learning model includes a one-dimensional convolutional neural network that can be configured to use a B allele frequency distribution as an input and encode it to an input size of 25 height, 1 width, and 100 depth. In some embodiments, each layer of the one-dimensional convolutional neural network performs a 1×1 convolution followed by a ReLU activation function.

[0023] In some embodiments, the trained machine learning model includes a two-dimensional convolutional neural network. In some embodiments, the B allele frequency distribution of the two-dimensional convolutional neural network is used as an input and encoded into an input size defined as height 25, width 100, and depth 1. In some embodiments, each convolutional layer of the two-dimensional convolutional neural network is followed by a ReLU layer. The output of the two-dimensional convolutional neural network may be a densely connected layer using a sigmoid activation function. Furthermore, hyperparameters of the two-dimensional convolutional neural network may be tuned by adjusting the number of layers, filter size, number of filters, or a combination thereof.

[0024] An estimated metric identifying tumor purity may be output. For example, a report including the estimated metric may be output. In some embodiments, the report includes information identifying a B allele frequency distribution. The report may also include information identifying at least one diagnostic marker and / or at least one prognostic marker. In some embodiments, the report includes information identifying predicted somatic variants. The report may also include a treatment recommendation. For example, the estimated tumor purity may indicate elevated methylation levels of O-6-methylguanine-DNA methyltransferase, a prognostic biomarker for glioblastoma. In another example, the estimated tumor purity can be used to measure the amount of mutational burden in cancer cells, which can be used to determine a particular type of immunotherapy. Depending on the amount of estimated tumor purity, a recommendation can be made as to whether to initiate treatment for glioblastoma. In some embodiments, the treatment recommendation includes a recommendation to administer treatment to the human subject. The treatment recommendation may include a recommendation not to administer treatment to the human subject.

[0025] Thus, embodiments of the present disclosure provide technical advantages over conventional systems by providing a more accurate estimate of tumor purity that does not require reliance on data from matched normal control samples. The estimated tumor purity can be used to improve the accuracy of analysis and annotation of sequencing data from tumor samples. One or more reports describing the estimated tumor purity can be generated (e.g., diagnostic and / or prognostic reports). For example, tumor purity estimates can be used to improve the accuracy of diagnostic techniques that identify somatic mutations and / or copy number alterations, and reports can be generated that include details of the predicted somatic mutations and / or copy number alterations.

[0026] Based on the tumor purity estimate and / or the report facilitated by the tumor purity estimate, one or more treatments can be administered to or withheld from the patient. For example, the predicted somatic variants can be compared to one or more databases of known cancer mutations to diagnose or characterize the cancer. Variants associated with responsiveness or non-responsiveness to particular cancer treatments can be identified, and treatment recommendations can be provided. The cancer can be treated based on the recommendations.

[0027] The following examples are provided to introduce particular embodiments. In the following description, for purposes of explanation, specific details are set forth to provide a thorough understanding of the examples in this disclosure. However, it will be apparent that various examples can be practiced without these specific details. For example, devices, systems, structures, assemblies, methods, and other components may be shown as components in block diagram form so as not to obscure the examples in unnecessary detail. In other examples, well-known devices, processes, systems, structures, and techniques may be shown without necessary detail so as to avoid obscuring the examples. The figures and descriptions are not limiting. The terms and expressions used in this disclosure are used in terms of description and not of limitation, and there is no intention to use such terms or expressions to exclude any equivalents of the features shown and described, or portions thereof. The word "example" is used herein to mean "serving as an example, instance, or illustration." Any embodiment or design described herein as an "example" is not necessarily to be construed as preferred or advantageous over other embodiments or designs.

[0028] II. Machine Learning Models for Estimating Tumor Purity from a Single Sample A. Example of a machine learning model for estimating tumor purity of a biological sample As described herein, the tumor purity of a biological sample can be estimated using a trained machine learning model. The trained machine learning model may correspond to one of various machine learning models trained to estimate the tumor purity of a biological sample. In some embodiments, the trained machine learning model includes multiple models (e.g., 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 machine learning models). For example, one of three machine learning models, such as a fully connected neural network, a one-dimensional convolutional neural network, or a two-dimensional convolutional neural network, can be trained to estimate tumor purity in nucleic acid sequencing data. In some cases, the trained machine learning model includes a deep neural network. Deep neural networks can be used to capture the internal structure of expanding, high-dimensional datasets (e.g., nucleic acid sequence data). Deep neural networks can identify high-level features, improving performance over traditional statistical models, increasing interpretability, and providing further understanding of the structure of nucleic acid sequence data.

[0029] A trained machine learning model may include hyperparameters. Hyperparameters may be external to the model and have values that are not estimated from data (e.g., training data or input data). In some cases, hyperparameters are tuned, e.g., to solve a particular predictive modeling problem. In some cases, hyperparameters are used to help estimate model parameters. Hyperparameters may be user-specified. In some cases, hyperparameters may be determined using a set of heuristic algorithms.

[0030] FIG. 1 shows a schematic diagram 100 of a fully connected neural network for estimating tumor purity according to some embodiments. The input layer 105 of the fully connected neural network is shown on the left, followed by a set of hidden layers 110. The output layer 115 is shown on the right. For a fully connected network, input features can include the whole-exome, B allele frequency (BAF) distribution of nucleotide sequence variants identified in the biological sample. The fully connected neural network may include a series of fully connected layers with a ReLU activation function, and the output activation function can be a sigmoid function. In some cases, the loss function of the fully connected neural network is configured to generate mean squared error (MSE). The fully connected neural network can be optimized by hyperparameter search using random sampling, such as a linear search for layers and size or a logarithmic search for the learning rate. Each output dimension of the fully connected neural network can depend on each input dimension. In some cases, the fully connected neural network can be a feedforward neural network.

[0031] Convolutional neural networks can be trained to estimate the tumor purity of biological samples. Convolutional neural networks rely on local connections and combined weights between units, followed by feature pooling (subsampling) to obtain translation-invariant descriptors. A basic convolutional neural network architecture includes one convolutional layer and a pooling layer, optionally followed by a fully connected layer for supervised prediction. In some cases, convolutional neural networks are configured with multiple (e.g., >10) convolutional and pooling layers to better model the input space. Convolutional neural networks may require large datasets to be adequately trained. In some embodiments, convolutional neural networks use fewer parameters than fully connected neural networks by computing convolutions in small regions of the input space and sharing parameters across regions. The convolutional neural network may be a one-dimensional convolutional neural network. The convolutional neural network may be a two-dimensional convolutional neural network. In some embodiments, the convolutional neural network has three or more dimensions.

[0032] FIG. 2 shows a schematic diagram 200 of a one-dimensional convolutional neural network for estimating tumor purity according to some embodiments. For the one-dimensional convolutional neural network, the input function may include a chromosome BAF distribution, which may be encoded into an input size defining a height of 25, a width of 1, and a depth of 100. Each layer of the one-dimensional convolutional neural network may perform a 1×1 convolution followed by a ReLU activation function. In some cases, a "network-within-a-network" deep network structure is used. In effect, one-dimensional (1D) convolutions with a flattened input and a stride of the depth of each layer are used. In some cases, hyperparameters corresponding to the depth and layer size are obtained from a fully connected neural network (see FIG. 1), and a hyperparameter corresponding to the learning rate of the one-dimensional convolutional neural network may be tuned for optimization.

[0033] FIG. 3 shows a schematic diagram 300 of a two-dimensional convolutional neural network for estimating tumor purity according to some embodiments. For the two-dimensional convolutional neural network, the input feature may include a chromosome BAF distribution and may be encoded into an input size defining a height of 25, a width of 100, and a depth of 1. Each layer corresponds to a two-dimensional convolutional layer (two-dimensional convolutional neural network layer) followed by a ReLU activation function. The output layer may correspond to a densely connected layer with a sigmoid activation function. In some embodiments, the hyperparameters of the two-dimensional convolutional neural network are optimized by searching for the number of layers, the filter size, and the number of filters. Dropout regularization and / or L2 regularization can be performed to reduce initial overfitting of the training set.

[0034] Other types of machine learning models can be trained to estimate the tumor purity of a biological sample. In some embodiments, the machine learning model corresponds to one or more of a gradient boosting decision tree (e.g., the XGBoost framework or the LightGBM framework), a bagging procedure, a boosting procedure, and / or a random forest algorithm. For example, a gradient boosted decision tree can be trained to estimate the tumor purity of a biological sample. Gradient boosting corresponds to a type of machine learning technique that can be used for regression and classification problems and to generate predictive models that may include an ensemble of weak predictive models, such as decision trees. In some cases, the gradient boosted decision tree can include, for example, the XGBoost framework or the LightGBM framework.

[0035] B. Training Dataset for Training Machine Learning Models to Estimate Tumor Purity of Biological Samples A machine learning model for estimating tumor purity of a biological sample can be trained using a supervised training algorithm. A supervised training algorithm may be used to train the machine learning model and generate outputs corresponding to linear and / or logistic regression values corresponding to tumor purity. The machine learning model can be trained based on a training dataset. In some cases, the training dataset includes sequencing datasets from pure tumor samples, pure normal samples, and a dilution series of tumor cells among normal cells. The sequencing datasets may be derived from biological samples of other subjects, including biological samples from subjects diagnosed with cancer. The biological samples in the training dataset may correspond to malignant tissue, benign tissue, or a mixture thereof. In some embodiments, the biological sample containing cancerous tissue is obtained without a matching normal sample. Additionally or alternatively, a matching normal sample is obtained for training and testing the machine learning model (e.g.,

[0036] In some embodiments, the training dataset includes a training nucleotide sequence variant dataset. The training variant dataset may correspond to nucleic acid sequence data in which nucleotide sequence variants are identified. Variants following a nucleotide may include single nucleotide polymorphisms (SNPs), one or more single nucleotide changes, insertion-deletion mutations (indels), small insertions, small deletions, structural variant junctions, variable length tandem repeats, flanking sequences, and combinations thereof. In some cases, biased "Variant Call Format (VCF)-level" datasets are generated using an artificial combination of VCF files generated from training samples, including tumor and normal samples. Thus, the training dataset may include in silico "dilutions" of tumor sequences. VCF-level datasets may be generated using the characteristics outlined in Table 1. To train a machine learning model, the machine learning model can be fitted to a "train" set, hyperparameters adjusted based on performance on the "train" and "dev" sets, and final performance evaluated on a "test" set. [Table 1]

[0037] In some embodiments, the training dataset includes training nucleic acid sequence data. The training nucleic acid sequence data may correspond to an unbiased "FASTQ-level" dataset generated by downsampling and combining reads from training samples. The training samples may include normal and tumor samples. FASTQ-level datasets can be generated with 99 and 11 unique samples and obtained by sequencing in vitro dilutions of tumor cells and matched normal cells. Compared to VCF-level datasets, FASTQ-level datasets are closer to the raw data because they may be at the read level rather than the variant level. In some cases, FASTQ-level datasets are used to improve or optimize performance using more realistic data. Because generating large FASTQ-level datasets is computationally challenging, machine learning models can be trained using FASTQ-level datasets using transfer learning. Transfer learning is performed using machine learning models pre-trained on VCF-level datasets, with 5-fold cross-validation. In some cases, each machine learning model can be trained on 4 / 5 of the training dataset and tested on 1 / 5 of the training dataset at each iteration.

[0038] C. Training and Testing Machine Learning Models to Estimate Tumor Purity of Biological Samples A machine learning model (e.g., a fully connected neural network) can be trained on a training dataset (e.g., a VCF-level dataset). In some cases, the machine learning model is initially trained on a VCF-level dataset. The machine learning model can be fit to a "train" set, hyperparameters can be adjusted based on performance on the "train" set and "dev" set, and final performance can be evaluated on a "test" set. FIG. 4A shows a loss vs. epoch plot 400A for a fully connected neural network, a one-dimensional convolutional neural network, and a two-dimensional convolutional neural network trained on a VCF-level dataset, according to some embodiments. The loss vs. epoch plot 400A shows that the performance level of each trained machine learning model increases with each epoch, thereby reaching an error value (e.g., mean squared error) in the range of approximately 0.01 to 0.025.

[0039] FIG. 4B shows root mean square error (RMSE) and mean absolute error (MAE) data 400B for a fully connected neural network, a one-dimensional convolutional neural network, and a two-dimensional convolutional neural network trained on a VCF-level dataset, according to some embodiments. All models perform comparably on the VCF-level dataset. For this dataset, root mean square error values range from 0.081 (two-dimensional convolutional neural network) to 0.088 (fully connected neural network) based on the training dataset. As shown in FIGS. 4A and 4B, the consistent performance of all machine learning models trained on the VCF-level training and test datasets indicates good fit and may be approaching optimality for estimating the tumor purity of a sample.

[0040] In some embodiments, the FASTQ level dataset is used to improve or optimize the performance of a trained machine learning model. Because generating a large FASTQ level dataset is computationally challenging, a machine learning model may be trained using the FASTQ level dataset using transfer learning. Transfer learning may be performed using a machine learning model pre-trained on the VCF level dataset. The performance of the trained machine learning model may be evaluated using a 5-fold cross-validation strategy. In some cases, each machine learning model may be trained on 4 / 5 of the training dataset and tested on 1 / 5 of the training dataset at each iteration. Figure 4C shows RSME and MAE data 400C for a fully connected neural network, a one-dimensional convolutional neural network, and a two-dimensional convolutional neural network trained on the FASTQ level dataset, according to some embodiments. The RMSE and MAE data 400C show comparable levels of performance between the trained one-dimensional convolutional neural network (e.g., 0.057 RMSE) and the trained two-dimensional convolutional neural network (e.g., 0.067 RMSE). Furthermore, the RMSE and MAE data 400C show that the trained machine learning model achieved low RMSE and MAE after transfer learning on the FASTQ-level dataset.

[0041] The performance level of each trained machine learning model was evaluated. Figure 5A shows a plot 500A comparing true tumor purity with that predicted by a trained fully connected network, according to some embodiments. Figure 5B shows a plot 500B comparing true tumor purity with that predicted by a trained two-dimensional convolutional neural network, according to some embodiments. Figure 5C shows a plot 500C comparing true tumor purity with that predicted by a one-dimensional convolutional neural network, according to some embodiments. The results from Figures 5A, 5B, and 5C, respectively, demonstrate that the trained machine learning models achieve performance levels comparable to existing tools that require matched normal control samples.

[0042] In some cases, convolutional neural network models tend to overfit to training datasets. For example, higher RSME and MAE values corresponding to convolutional neural networks may not be significantly improved by regularization and dropout. The tendency of these convolutional neural networks to overfit may indicate the ability of these machine learning models to detect complex signals of tumor purity when trained using larger training datasets. Therefore, machine learning models may be further trained using larger and more diverse training datasets.

[0043] In some embodiments, the trained machine learning model is trained using the training data and / or test data to achieve one or more predetermined performance levels for estimating tumor purity. The one or more predetermined performance levels include: An MAE of up to 1%, 2%, 3%, 4%, 5%, 6%, 7%, 8%, 9%, 10%, 11%, 12%, 13%, 14%, 15%, 16%, 17%, 18%, 19%, or 20%. RMSE of up to 1%, 2%, 3%, 4%, 5%, 6%, 7%, 8%, 9%, 10%, 11%, 12%, 13%, 14%, 15%, 16%, 17%, 18%, 19%, or 20%. Accuracy of at least about 0.2, 0.25, 0.3, 0.35, 0.4, 0.45, 0.5, 0.55, 0.6, 0.65, 0.7, 0.75, 0.8, 0.85, 0.9, 0.95, or higher. Additionally or alternatively, the trained machine learning model has an accuracy of about 0.2-1.0, 0.2-0.9, 0.2-0.8, 0.2-0.7, 0.2-0.6, 0.2-0.5, 0.2-0.4, 0.2-0.3, 0.3-1.0, 0.3-0.9, 0.3-0.8, 0.3-0.7, 0.3-0.6, 0.3-0.5, 0.3-0.4, 0.4-1.0, 0.4-0.9, 0.4-0. It is trained to estimate tumor purity with an accuracy of 8, 0.4–0.7, 0.4–0.6, 0.4–0.5, 0.5–1.0, 0.5–0.9, 0.5–0.8, 0.5–0.7, 0.5–0.6, 0.6–1.0, 0.6–0.9, 0.6–0.8, 0.6–0.7, 0.7–1.0, 0.7–0.9, 0.7–0.8, 0.8–1.0, 0.8–0.9, or 0.9–1.0. Additionally or alternatively, the trained machine learning model has a recall of at least about 0.2, 0.25, 0.3, 0.35, 0.4, 0.45, 0.5, 0.55, 0.6, 0.65, 0.7, 0.75, 0.8, 0.85, 0.9, 0.95, or more. Additionally or alternatively, the trained machine learning model has a recall of at least about 0.2-1.0, 0.2-0.9, 0.2-0.8, 0.2-0.7, 0.2-0.6, 0.2-0.5, 0.2-0.4, 0.2-0.3, 0.3-1.0, 0.3-0.9, 0.3-0.8, 0.3-0.7, 0.3-0.6, 0.3-0.5, 0.3-0.4, 0.4-1.0, 0.4-0.9, 0.4-0.8 , trained to estimate tumor purity with a recall of 0.4–0.7, 0.4–0.6, 0.4–0.5, 0.5–1.0, 0.5–0.9, 0.5–0.8, 0.5–0.7, 0.5–0.6, 0.6–1.0, 0.6–0.9, 0.6–0.8, 0.6–0.7, 0.7–1.0, 0.7–0.9, 0.7–0.8, 0.8–1.0, 0.8–0.9, or 0.9–1.0. An F1 score (e.g., macro-average F1 classification score) of at least about 0.2, 0.25, 0.3, 0.35, 0.4, 0.45, 0.5, 0.55, 0.6, 0.65, 0.7, 0.75, 0.8, 0.85, 0.86, 0.87, 0.88, 0.89, 0.9, 0.91, 0.92, 0.93, 0.94, 0.95, 0.96, 0.97, 0.98, 0.99, 0.995, or higher. Additionally or alternatively, the trained machine learning model may be selected from the group consisting of approximately 0.2-1.0, 0.2-0.99, 0.2-0.95, 0.2-0.9, 0.2-0.8, 0.2-0.7, 0.2-0.6, 0.2-0.5, 0.2-0.4, 0.2-0.3, 0.3-1.0, 0.3-0.99, 0.2-0.95, 0.3-0. 9, 0.3~0.8, 0.3~0.7, 0.3~0.6, 0.3~0.5, 0.3~0.4, 0.4~1.0, 0.4~0.99, 0.4~0.95, 0.4~0.9, 0.4~0.8, 0.4~0.7, 0.4~0.6, 0.4~0.5, 0.5~1.0, 0.5~0.99, 0.5~0.95, 0.5~0.9, 0.5~0.8, 0.5~0.7, 0.5~0.6, 0.6~1.0, 0.6~0.99, 0.6~0.95, 0.6~0.9, 0.6~0.8, 0.6~0.7, 0.7~1.0, 0.7~0.99, 0.7~0.98, 0.7~0.97, 0.7~0.96, 0.7~0.95, 0.7~0.9, 0.7~0. It is trained to estimate tumor purity with an F1 score of 8, 0.8–1.0, 0.8–0.99, 0.8–0.98, 0.8–0.97, 0.8–0.96, 0.8–0.95, 0.8–0.9, 0.9–1.0, 0.9–0.99, 0.9–0.98, 0.9–0.97, 0.9–0.96, or 0.9–0.95.

[0044] III. Estimation of Tumor Purity 6 shows an example schematic 600 for estimating tumor purity using a trained machine learning model, according to some embodiments. Tumor purity from sequencing of unmatched biological samples can be estimated using a trained machine learning model (e.g., a regression model). In some cases, a deep neural network is used to estimate tumor purity from the allele fraction distribution (e.g., B allele frequency distribution) of nucleotide sequence variants in unpaired biological samples. A regression model can be used to predict tumor purity from the B allele frequency (BAF) of whole exome sequencing data.

[0045] In block 605, a mismatched tumor sample can be obtained from a cancer patient (i.e., without a matched normal sample). In block 610, DNA is extracted from the tumor sample, processed, and whole-exome sequenced. Optionally, the sequencing reads undergo quality control processing (e.g., via FastQC) to provide a FASTQ file. The FASTQ file is aligned to a reference genome to generate a BAM file.

[0046] In block 615, the GATK HaplotypeCaller can be used to call variants from the BAM file and generate a VCF file containing variant information. Heterozygous sites within the sample exome can be identified, and the VCF file can be filtered to obtain reference and alternate read depths for all heterozygous sites. This information is used to calculate B allele frequency (BAF). A normalized BAF distribution is calculated for heterozygous sites across the exome.

[0047] In block 620, the trained machine learning model can be used to estimate tumor purity from the normalized BAF distribution. Thus, the trained machine learning model can generate accurate estimates of tumor purity even in the absence of a matched normal sample.

[0048] A. Subjects and Samples To estimate tumor purity, nucleic acid sequence data representing multiple nucleic acid molecules can be obtained from a subject's biological sample. The subject can be human. The subject can be male or female. The subject can be a fetus, infant, child, adolescent, teenager, or adult. The subject can be a patient of any age. For example, the subject can be a patient under about 10 years of age. For example, the subject can be a patient at least about 0, 5, 10, 20, 30, 40, 50, 60, 70, 80, 90, or 100 years of age. The subject can be a patient or other individual receiving or being evaluated for a therapeutic regimen (e.g., cancer treatment). However, in some cases, the subject is not receiving a therapeutic regimen.

[0049] In some cases, the subject may be a mammal or a non-mammal. In some cases, the subject is a mammal, such as a human, a non-human primate (e.g., ape, monkey, chimpanzee), cat, dog, rabbit, goat, horse, cow, pig, rodent, mouse, SCID mouse, rat, guinea pig, or sheep. In some embodiments, species variants or homologs of these genes are used in non-human animal models. Species variants can be genes from different species that have the greatest sequence identity and similarity in functional properties to each other. Many such species variants of human genes may be listed in the Swiss-Prot database.

[0050] Certain embodiments may include obtaining a sample from a subject, such as a human subject. In some cases, a clinical specimen from the patient is obtained. For example, blood may be drawn from the patient. Certain embodiments may include specifically detecting, profiling, or quantifying molecules (e.g., nucleic acids, DNA, RNA, etc.) within the biological sample.

[0051] The sample may be a tissue sample or a bodily fluid. Optionally, the sample is an organ sample, such as a tissue sample or biopsy. Optionally, the sample contains cancer cells. Optionally, the sample contains cancerous cells and normal cells. Optionally, the sample is a tumor biopsy. The bodily fluid may be sweat, saliva, tears, urine, blood, menstrual fluid, semen, and / or cerebrospinal fluid. Optionally, the sample is a blood sample. The sample may contain one or more peripheral blood lymphocytes. The sample may be a whole blood sample. The blood sample may be a peripheral blood sample. Optionally, the sample contains peripheral blood mononuclear cells (PBMCs), and optionally, the sample contains peripheral blood lymphocytes (PBLs). The sample may be a serum sample.

[0052] The sample can be obtained using any method capable of providing a sample suitable for the analytical methods described herein. The sample may be obtained by a non-invasive method, such as a throat swab, oral swab, bronchial wash, urine collection, skin or cervical scrape, buccal swab, saliva collection, stool collection, menstrual blood collection, or semen collection. The sample may be obtained by a minimally invasive method, such as a blood draw. The sample may be obtained by venipuncture. In other examples, the sample is obtained by an invasive procedure, including, but not limited to, a biopsy, alveolar or pulmonary lavage, or needle aspiration. The biopsy method may include a surgical biopsy, an incisional biopsy, an excisional biopsy, a punch biopsy, a shaved biopsy, or a skin biopsy. The sample may be a formalin-fixed section. The needle aspiration method may further include a fine needle aspiration, a core needle biopsy, a vacuum-assisted biopsy, or a large core biopsy. In some cases, multiple samples may be obtained by the methods herein to ensure a sufficient amount of biological material. In some cases, the sample is not obtained by biopsy. Optionally, the sample is not a kidney biopsy.

[0053] The methods of the present disclosure can be used to estimate tumor purity in samples containing at least about 1%, 5%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 95%, or more tumor cells. In some embodiments, the methods of the present disclosure can be used to estimate tumor purity in samples containing up to about 10%, 15%, 20%, 25%, 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 95%, or more tumor cells.

[0054] B. Generating Nucleic Acid Sequencing Data In some embodiments, the sample is processed to obtain nucleic acid sequence data. A "nucleic acid" or "nucleic acid molecule" can refer to a polymeric form of nucleotides of any length, either ribonucleotides, deoxyribonucleotides, or peptide nucleic acids (PNAs), containing purine and pyrimidine bases or other natural, chemically or biochemically modified, non-natural, or derivatized nucleotide bases. The backbone of a polynucleotide can contain sugars and phosphate groups typically found in RNA or DNA, or modified or substituted sugars or phosphate groups. A polynucleotide can also contain modified nucleotides, such as methylated nucleotides and nucleotide analogs. The sequence of nucleotides can be interrupted by non-nucleotide components. Thus, the terms nucleoside, nucleotide, deoxynucleoside, and deoxynucleotide generally include analogs, such as those described herein. These analogs are molecules that share some structural features with naturally occurring nucleosides or nucleotides, and thus, when incorporated into a nucleic acid or oligonucleoside sequence, allow hybridization with naturally occurring nucleic acid sequences in solution. Typically, these analogs are derived from naturally occurring nucleosides and nucleotides by substituting and / or modifying the base, ribose, or phosphodiester moiety. This change can be tailored to stabilize or destabilize hybridization or, as needed, to increase the specificity of hybridization with complementary nucleic acid sequences. The nucleic acid molecule may be a DNA molecule. The nucleic acid molecule may be an RNA molecule.

[0055] Sample processing includes nucleic acid sample processing followed by nucleic acid sample sequencing. Part or all of the biological sample can be sequenced to provide nucleic acid sequence data, which can be stored or maintained in electronic, magnetic, or optical storage. The sequence information can be analyzed by a computer processor, and the analyzed sequence information can be stored in an electronic storage location. The electronic storage location can contain a pool or collection of sequence information generated from the nucleic acid sample and analyzed sequence information. In some embodiments, the biological sample is obtained from a subject with or suspected of having cancer.

[0056] In some embodiments, nucleic acid sequencing data is generated from pure tumor and pure normal samples. Matched paired cell lines are available from separate sources (e.g., the American Type Culture Collection (ATCC)). Each matched pair may include a tumor cell line and a normal cell line from the same subject. The cell lines may be cultured and expanded in vitro to obtain a suitable number of cells for DNA extraction. DNA is extracted, processed, and subjected to whole-exome or whole-genome sequencing. Sequence reads may undergo quality control processing (e.g., via FastQC) to provide FASTQ files.

[0057] In some cases, the nucleic acid sequence data is generated using whole genome sequencing. In some cases, whole genome sequencing is used to identify variants in an individual. In some cases, sequencing can include deep sequencing across a portion of the genome. For example, a fraction of the genome can be at least about 50, 75, 100, 125, 150, 175, 200, 225, 250, 275, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1,000, 1100, 1200, 1300, 1400, 1500, 1600, 1700, 1800, 2100, 2250, 2300, 2400, 2500, 2600, 2750, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1,000, 1100, 1200, 1300, 1400, 1500, 1600, 1700, 1800, 2100, 2250, 2500, 2750, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1,000, 1100, 1 The genome may be sequenced over 800, 1900, 2,000, 3,000, 4,000, 5,000, 6,000, 7,000, 8,000, 9,000, 10,000, 15,000, 20,000, 30,000, 40,000, 50,000, 60,000, 70,000, 80,000, 90,000, 100,000, or more bases or base pairs. In some cases, the genome may be sequenced over 1 million, 2 million, 3 million, 4 million, 5 million, 6 million, 7 million, 8 million, 9 million, 10 million, or more than 10 million bases or base pairs. In some cases, the genome may be sequenced over the entire exome (e.g., whole exome sequencing). In some cases, deep sequencing can include obtaining multiple reads across a portion of the genome. For example, obtaining multiple reads can include at least 2, 3, 4, 5, 6, 7, 8, 9, 10, 20, 30, 40, 50, 60, 70, 80, 90, 100, 200, 300, 400, 500, 600, 700, 800, 900, 1000, 10,000 or more reads across a portion of the genome.

[0058] In some cases, generating nucleic acid sequence data includes detecting low allele fractions by deep sequencing. In some cases, deep sequencing is performed by next-generation sequencing. In some cases, deep sequencing is performed by avoiding error-prone regions. In some cases, error-prone regions may include regions near sequence duplications, regions with abnormally high or low %GC, regions near homopolymers, dinucleotides and trinucleotides, and regions near other short repeats. In some cases, error-prone regions may include regions that lead to DNA sequencing errors (e.g., polymerase slippage in homopolymer sequences).

[0059] In some cases, generating nucleic acid sequence data includes performing one or more sequencing reactions on one or more nucleic acid molecules in the sample. Certain embodiments may include performing 1 or more, 2 or more, 3 or more, 4 or more, 5 or more, 6 or more, 7 or more, 8 or more, 9 or more, 10 or more, 15 or more, 20 or more, 30 or more, 40 or more, 50 or more, 60 or more, 70 or more, 80 or more, 90 or more, 100 or more, 200 or more, 300 or more, 400 or more, 500 or more, 600 or more, 700 or more, 800 or more, 900 or more, or 1000 or more sequencing reactions on one or more nucleic acid molecules in the sample. The sequencing reactions may be performed simultaneously, sequentially, or a combination thereof. The sequencing reactions may include whole genome sequencing or exome sequencing. The sequencing reactions may include a Maxim-Gilbert system, a chain termination system, or a high-throughput system. Alternatively or additionally, the sequencing reaction may include HeliScope™ single molecule sequencing, nanopore DNA sequencing, Lynx Therapeutics' Massively Parallel Signature Sequencing (MPSS), 454 pyrosequencing, single molecule real-time (RNAP) sequencing, Illumina (Solexa) sequencing, SOLiD sequencing, Ion Torrent™ ion semiconductor sequencing, single molecule SMRT™ sequencing, polony sequencing, DNA nanoball sequencing, an approach by VisiGen Biotechnologies, or a combination thereof.Alternatively or additionally, the sequencing reaction may include one or more sequencing platforms, such as, but not limited to, single molecule real time (SMRT™) technology such as the Illumina Genome Analyzer IIx, HiSeq, and MiSeq, the PacBio RS system from Pacific Biosciences (California), and true single molecule sequencing (tSMS™) technology such as the Solexa sequencer (Illumina), and the HeliScope™ sequencer from Helicos (Cambridge, Massachusetts). The sequencing reaction may also include an electron microscope or a chemically sensitive field effect transistor (chemFET) array. In some embodiments of the present disclosure, the sequencing reaction comprises capillary sequencing, next-generation sequencing, Sanger sequencing, sequencing by synthesis, sequencing by ligation, sequencing by hybridization, single molecule sequencing, or a combination thereof. Sequencing by synthesis may comprise reversible terminator sequencing, processive single molecule sequencing, sequence flow sequencing, or a combination thereof. Sequence flow sequencing may comprise pyrosequencing, pH-mediated sequencing, semiconductor sequencing, or a combination thereof.

[0060] In some cases, generating the nucleic acid sequence data includes performing at least one long read sequencing reaction and at least one short read sequencing reaction. The long read sequencing reaction and / or the short read sequencing reaction may be performed on at least a portion of a subset of the nucleic acid molecules. The long read sequencing reaction and / or the short read sequencing reaction may be performed on at least a portion of two or more subsets of the nucleic acid molecules. Both the long read sequencing reaction and the short read sequencing reaction may be performed on at least a portion of one or more subsets of the nucleic acid molecules.

[0061] Sequencing of one or more nucleic acid molecules or a subset thereof may be at least about 5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 60, 70, 80, 90, 100, 200, 300, 400, 500, 600, 700, 800, 900, 1,000, 1500, 2,000, 2500, 3,000, 3500, 4,000, 4500, 5,000, 5500, 6,000, 6500, 7,000, 7500, 8,000, 9,000, 10,000, 11,000, 12,000, 13,000, 14,000, 15,000, 16,000, 17,000, 18,000, 19,000, 21,000, 22,000, 23,000, 24,000, 25,000, 26,000, 27,000, 28,000, 29,000, 30,000, 31,000, 32,000, 33,000, 34,000, 35,000, 40,000, 45,000, 50,000, 55,000, 60,000, 65,000, 70,000, 75,000, 80,000, 90,000, 100,00 The sequence may comprise 00, 8500, 9,000, 10,000, 25,000, 50,000, 75,000, 100,000, 250,000, 500,000, 750,000, 10,000,000, 25,000,000, 50,000,000, 100,000,000, 250,000,000, 500,000,000, 750,000,000, 1,000,000,000 or more sequencing reads.

[0062] The sequencing reaction may comprise at least about 50, 60, 70, 80, 90, 100, 110, 120, 130, 140, 150, 160, 170, 180, 190, 200, 210, 220, 230, 240, 250, 260, 270, 280, 290, 300, 325, 350, 375, 400, 425, 450, 475, 500, 600, 700, 800, 900, 1,000, 1,500, 2,600, 3,700, 4,800, 5,900, 6,100, 7,100, 8,100, 9,100, 10,150, 11,100, 12,100, 13,100, 14,100, 15,100, 16,100, 17,100, 18,100, 19,100, 20,110, 210, 220, 230, 240, 250, 260, 270, 280, 290, 300, 325, 350, 375, 400, 425, 450, 475, 500, 600, 700, 800, 900, 1,000, 15,100, 16,100, 17,100, 18,100, 19,100, 20 This may include sequencing 2,000, 2500, 3,000, 3500, 4,000, 4500, 5,000, 5500, 6,000, 6500, 7,000, 7500, 8,000, 8500, 9,000, 10,000, 20,000, 30,000, 40,000, 50,000, 60,000, 70,000, 80,000, 90,000, 100,000 or more bases or base pairs. The sequencing reaction may comprise at least about 50, 60, 70, 80, 90, 100, 110, 120, 130, 140, 150, 160, 170, 180, 190, 200, 210, 220, 230, 240, 250, 260, 270, 280, 290, 300, 325, 350, 375, 400, 425, 450, 475, 500, 600, 700, 800, 900, 1000, 1500, 2000, 3100, 3250, 3300, 3400, 3500, 3600, 3750, 4000, 4250, 4500, 4750, 5000, 6000, 7000, 8000, 9000, 10000, 15000, 16000, 17000, 18000, 19000, 20000, 21000, 22000, 23000, 24000, 25000, 26000, 27000, 28000, 29000, 30000, 32500, 35000, 37500, 40000, 42500, 45000, 47500, 50000, 60000, 70000, 80000, 90000, The method may include sequencing 000, 2500, 3,000, 3500, 4,000, 4500, 5,000, 5500, 6,000, 6500, 7,000, 7500, 8,000, 8500, 9,000, 10,000, 20,000, 30,000, 40,000, 50,000, 60,000, 70,000, 80,000, 90,000, 100,000 or more consecutive bases or base pairs.

[0063] In some cases, the sequencing technique generates at least 100 reads / run (reads / run), at least 200 reads / run, at least 300 reads / run, at least 400 reads / run, at least 500 reads / run, at least 600 reads / run, at least 700 reads / run, at least 800 reads / run, at least 900 reads / run, at least 1000 reads / run, at least 5,000 reads / run, at least 10,000 reads / run, at least 50,000 reads / run, at least 100,000 reads / run, at least 500,000 reads / run, or at least 1,000,000 reads / run. Alternatively, the sequencing technique generates at least 1,500,000 reads / run, at least 2,000,000 reads / run, at least 2,500,000 reads / run, at least 3,000,000 reads / run, at least 3,500,000 reads / run, at least 4,000,000 reads / run, at least 4,500,000 reads / run, or at least 5,000,000 reads / run.

[0064] In some cases, the sequencing approach generates at least about 30 base pairs, at least about 40 base pairs, at least about 50 base pairs, at least about 60 base pairs, at least about 70 base pairs, at least about 80 base pairs, at least about 90 base pairs, at least about 100 base pairs, at least about 110, at least about 120 base pairs per read, at least about 150 base pairs, at least about 200 base pairs, at least about 250 base pairs, at least about 300 base pairs, at least about 350 base pairs, at least about 400 base pairs, at least about 450 base pairs, at least about 500 base pairs, at least about 550 base pairs, at least about 600 base pairs, at least about 700 base pairs, at least about 800 base pairs, at least about 900 base pairs, or at least about 1,000 base pairs per read. Additionally or alternatively, sequencing techniques may generate long sequencing reads. In some cases, the sequencing technique provides a sequence of at least about 1,200 base pairs / read, at least about 1,500 base pairs / read, at least about 1,800 base pairs / read, at least about 2,000 base pairs / read, at least about 2,500 base pairs / read, at least about 3,000 base pairs / read, at least about 3,500 base pairs / read, at least about 4,000 base pairs / read, at least about 4,500 base pairs / read, at least about 5,000 base pairs / read, at least about 6,000 base pairs / read, pairs / read, at least about 7,000 base pairs / read, at least about 8,000 base pairs / read, at least about 9,000 base pairs / read, at least about 10,000 base pairs / read, 20,000 base pairs / read, 30,000 base pairs / read, 40,000 base pairs / read, 50,000 base pairs / read, 60,000 base pairs / read, 70,000 base pairs / read, 80,000 base pairs / read, 90,000 base pairs / read, or 100,000 base pairs / read.

[0065] High-throughput sequencing systems can allow for detection of sequenced nucleotides immediately after or as they are incorporated into the growing strand, i.e., detection of sequences in real time or substantially real time. In some cases, high-throughput sequencing generates at least 1,000, at least 5,000, at least 10,000, at least 20,000, at least 30,000, at least 40,000, at least 50,000, at least 100,000, or at least 500,000 sequence reads per hour, with each read being at least 50, at least 60, at least 70, at least 80, at least 90, at least 100, at least 120, at least 150, at least 200, at least 250, at least 300, at least 350, at least 400, at least 450, or at least 500 bases per read. Sequencing can be performed using nucleic acids described herein, such as genomic DNA, cDNA derived from RNA transcripts, or RNA as a template.

[0066] C. Identification of Nucleotide Sequence Variants To estimate the tumor purity of a biological sample, the B allele frequency (BAF) of the biological sample can be determined. To determine the BAF, a set of genomic regions can be identified, and each of the set of genomic regions can contain nucleotide sequence variants relative to the corresponding genomic region of a reference genome. To determine the set of genomic regions, nucleic acid sequence data can be aligned to a reference genome. For example, a FASTQ file corresponding to the nucleic acid sequence data can be aligned to a reference genome to generate one or more BAM files. The one or more BAM files can be processed by another module (e.g., GATK HaplotypeCaller) to identify the set of genomic regions. In some cases, a VCF file containing the set of genomic regions is generated. Additionally or alternatively, a VCF file can be obtained for a matched paired cell line. The VCF file can identify a set of genomic regions with nucleotide sequence variants based on a comparison of the nucleic acid sequence data with a reference genome.

[0067] Certain embodiments may include nucleic acid molecules comprising one or more genomic regions. Certain embodiments may include nucleic acid molecules comprising one or more sets of genomic regions. The one or more genomic regions may comprise one or more genomic region features. The genomic region features may comprise an entire genome or a portion thereof. The genomic region features may comprise an entire exome or a portion thereof. The genomic region features may comprise one or more gene sets. The genomic region features may comprise one or more genes. The genomic region features may comprise one or more sets of regulatory elements. The genomic region features may comprise one or more regulatory elements.

[0068] The genomic region features may include a set of polymorphisms. The genomic region features may include one or more polymorphisms. The genomic region features may be related to the GC content, complexity, and / or mappability of one or more nucleic acid molecules. The genomic region features may include one or more simple tandem repeats (STRs), unstable expansion repeats, segmental duplications, single and paired read degenerate mapping scores, GRCh37 patches, or a combination thereof. The genomic region features may include one or more low average coverage regions from whole genome sequencing (WGS), zero average coverage regions from WGS, validated compaction, or a combination thereof. The genomic region features may include one or more alternative or non-reference sequences. The genomic region features may include one or more gene phasing and reconstituted genes. In some aspects of the present disclosure, one or more genomic region features are not mutually exclusive. For example, a feature of a genomic region comprising an entire genome or a portion thereof may overlap with a feature of an additional genomic region, such as an entire exome or a portion thereof, one or more genes, one or more regulatory elements, etc. Alternatively, one or more features of a genomic region may be mutually exclusive. For example, a genomic region comprising a non-coding portion of an entire genome may not overlap with a feature of a genomic region, such as an exome or a portion thereof, or a coding portion of a gene. Alternatively or additionally, one or more features of a genomic region may be partially exclusive or partially inclusive. For example, a genomic region comprising an entire exon or a portion thereof may partially overlap with a genomic region comprising an exon portion of a gene. However, a genomic region comprising an entire exome or a portion thereof may not overlap with a genomic region comprising an intron portion of a gene. Thus, a feature of a genomic region comprising a gene or a portion thereof may partially exclude and / or partially include a feature of a genomic region comprising an entire exome or a portion thereof.

[0069] Certain embodiments may include a nucleic acid sample or molecule comprising one or more genomic regions, at least one of which comprises a genomic region characteristic of an entire genome or a portion thereof. The entire genome or a portion thereof may comprise one or more coding portions of the genome, one or more non-coding portions of the genome, or a combination thereof. The coding portion of the genome may comprise one or more coding portions of genes encoding one or more proteins. The one or more coding portions of the genome may comprise an entire exome or a portion thereof. Alternatively or additionally, the one or more coding portions of the genome may comprise one or more exons.

[0070] The one or more non-coding portions of the genome may include one or more non-coding molecules or portions thereof. The non-coding molecules may include one or more non-coding RNAs, one or more regulatory elements, one or more introns, one or more pseudogenes, one or more repeat sequences, one or more transposons, one or more viral elements, one or more telomeres, portions thereof, or combinations thereof. The non-coding RNA may be a functional RNA molecule that is not translated into protein. Examples of non-coding RNA include, but are not limited to, ribosomal RNA, transfer RNA, PIWI-interacting RNA, microRNA, siRNA, shRNA, snoRNA, sncRNA, and lncRNA. Pseudogenes may be associated with known genes and are typically no longer expressed. The repeat sequence may include one or more tandem repeats, one or more interspersed repeats, or a combination thereof. The tandem repeat may include one or more satellite DNAs, one or more minisatellites, one or more microsatellites, or a combination thereof.

[0071] Interspersed repeats may include one or more transposons. Transposons may be mobile genetic elements. Mobile genetic elements often have variable locations within the genome. Transposons can be classified as class I transposons (class I TEs) or class II transposons (class II TEs). Class I TEs (e.g., retrotransposons) often copy themselves in two steps: first from DNA to RNA by transcription, and then from RNA back to DNA by reverse transcription. The DNA copy can then be inserted into a new location in the genome. Class I TEs may include one or more long terminal repeats (LTRs), one or more long interspersed nuclear elements (LINEs), one or more short interspersed nuclear elements (SINEs), or a combination thereof. Examples of LTRs include, but are not limited to, human endogenous retroviruses (HERVs), medium reiterated repeats 4 (MER4), and retrotransposons. Examples of LINEs include, but are not limited to, LINE1 and LINE2. SINEs may include one or more Alu sequences, one or more mammalian-wide interspersed repeats (MIRs), or a combination thereof. Class II TEs (e.g., DNA transposons) often do not involve an RNA intermediate. DNA transposons are often excised from one site and inserted into another site in the genome. Alternatively, DNA transposons replicate and insert into a new location in the genome. Examples of DNA transposons include, but are not limited to, MER1, MER2, and mariner. Viral elements may include one or more endogenous retroviral sequences. Telomeres are often regions of repetitive DNA at the ends of chromosomes.

[0072] Certain embodiments may include a subset of nucleic acid samples or nucleic acid molecules comprising one or more genomic regions, at least one of which comprises a genomic region characteristic of an entire exome or a portion thereof. An exome is often a portion of a genome formed by exons. An exome may be formed by untranslated regions (UTRs), splice sites, and / or intronic regions. An entire exome or a portion thereof may include one or more exons of a protein-coding gene. An entire exome or a portion thereof may include one or more untranslated regions (UTRs), splice sites, and introns.

[0073] Certain embodiments may include a nucleic acid sample or molecule comprising one or more genomic regions, at least one of which comprises a genomic region characteristic of a gene or a portion thereof. Typically, a gene comprises a stretch of nucleic acid encoding a polypeptide or functional RNA. A gene may comprise one or more exons, one or more introns, one or more untranslated regions (UTRs), or a combination thereof. Exons often encode sections of a gene, are transcribed into precursor mRNA sequences, and are present in the gene's final mature RNA product. Introns often represent non-coding sections of a gene, are transcribed into precursor mRNA sequences, and are removed by RNA splicing. UTRs may refer to the sections on either side of the coding sequence of an mRNA strand. A UTR located 5' from a coding sequence may be referred to as a 5'UTR (or leader sequence). A UTR located 3' from a coding sequence may be referred to as a 3'UTR (or trailer sequence). A UTR may contain one or more elements for regulating gene expression. Elements such as regulatory elements may be located in the 5'UTR. Regulatory sequences such as polyadenylation signals, protein binding sites, and miRNA binding sites can be located in the 3'UTR. Protein binding sites located in the 3'UTR include, but are not limited to, selenocysteine insertion sequence (SECIS) elements and AU-rich elements (AREs). SECIS elements may instruct ribosomes to translate the codon UGA as selenocysteine rather than as a stop codon. AREs are often stretches primarily composed of adenine and uracil nucleotides, which may affect mRNA stability.

[0074] Certain embodiments may include a subset of nucleic acid samples or nucleic acid molecules comprising one or more genomic regions, at least one of which comprises a genomic region characteristic of a set of genes. The set of genes may include, but is not limited to, Mendel DB genes, Human Gene Mutation Database (HGMD) genes, Cancer Gene Census genes, Online Mendelian Inheritance in Man (OMIM) genes, HGMD Mendelian genes, and human leukocyte antigen (HLA) genes. The set of genes may have one or more known Mendelian traits, one or more known disease traits, one or more known drug traits, one or more known biomedically interpretable variants, or a combination thereof. A Mendelian trait may be controlled by a single genetic locus and exhibit a Mendelian inheritance pattern. A set of genes with known Mendelian traits may include one or more genes encoding Mendelian traits including, but not limited to, taste for phenylthiocarbamide (dominant), smell for hydrogen cyanide (bitter almond-like) (recessive), albinism (recessive), brachydactyly (short fingers and toes), and wet (dominant) or dry (recessive) earwax. Disease traits may cause or increase the risk of disease and may be inherited in Mendelian or complex patterns. A set of genes with known disease characteristics may include one or more genes encoding disease traits including, but not limited to, cystic fibrosis, hemophilia, and Lynch syndrome.

[0075] A drug trait may alter the metabolism, optimal dose, adverse reactions, and side effects of one or more drugs or drug families. The set of genes with known drug traits may include one or more genes encoding drug traits, including, but not limited to, CYP2D6, UGT1A1, and ADRB1. The biomedically interpretable variant may be a genetic polymorphism associated with a disease or indication. The set of genes with known biomedically interpretable variants may include one or more genes encoding biomedically interpretable variants, including, but not limited to, cystic fibrosis (CF) mutations, muscular dystrophy mutations, p53 mutations, Rb mutations, cell cycle regulators, receptors, and kinases. Alternatively or additionally, the set of genes with known biomedically interpretable variants may include one or more genes associated with Huntington's disease, cancer, cystic fibrosis, or muscular dystrophy (e.g., Duchenne muscular dystrophy).

[0076] Certain embodiments may include nucleic acid samples or molecules comprising one or more genomic regions, at least one of which comprises a genomic region characteristic of a regulatory element or portion thereof. Regulatory elements may be cis-regulatory elements or trans-regulatory elements. Cis-regulatory elements may be sequences that control the transcription of nearby genes. Cis-regulatory elements may be located within 5' or 3' untranslated regions (UTRs) or introns. Trans-regulatory elements may control the transcription of distant genes. Regulatory elements may include one or more promoters, one or more enhancers, or a combination thereof. Promoters may enhance the transcription of a particular gene and may be found upstream of the coding region. Enhancers may exert a distant effect on the transcription level of a gene.

[0077] Certain embodiments may include a subset of nucleic acid samples or molecules comprising one or more genomic regions, at least one of which comprises a genomic region characteristic of the region, including a polymorphism or a portion thereof. Generally, a polymorphism refers to a genotypic mutation. A polymorphism may be a germline variant or a somatic variant. A polymorphism may comprise one or more base changes, insertions, repeats, or deletions of one or more bases. Copy number variations (CNVs), transversions, and other rearrangements are also forms of genetic variation. Polymorphic markers include restriction fragment length polymorphisms, variable number tandem repeats (VNTRs), hypervariable regions, minisatellites, dinucleotide repeats, trinucleotide repeats, tetranucleotide repeats, simple sequence repeats, and insertion elements such as Alu. The allelic form occurring most frequently in a selected population may be referred to as the wild-type form. Diploid organisms may be homozygous or heterozygous for an allelic form. There are two forms of diallelic polymorphisms. There are three forms of triallelic polymorphisms. A single nucleotide polymorphism (SNP) is one form of polymorphism. In some aspects of the present disclosure, the one or more polymorphisms include one or more single nucleotide changes, indels, small insertions, small deletions, structural variant junctions, variable length tandem repeats, flanking sequences, or combinations thereof. The one or more polymorphisms may be located within coding and / or non-coding regions. The one or more polymorphisms may be located within, around, or near genes, exons, introns, splice sites, untranslated regions, or combinations thereof. The one or more polymorphisms may span at least a portion of a gene, exon, intron, or untranslated region.

[0078] Certain embodiments may include a nucleic acid sample or molecule comprising one or more genomic regions, at least one of which comprises a genomic region feature comprising one or more simple tandem repeats (STRs), unstable expansion repeats, segmental duplications, single and paired read degenerate mapping scores, GRCh37 patches, or a combination thereof. The one or more STRs may comprise one or more homopolymers, one or more dinucleotide repeats, one or more trinucleotide repeats, or a combination thereof. The one or more homopolymers may be about 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, or more bases or base pairs. The dinucleotide and / or trinucleotide repeats can be about 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 35, 40, 45, 50 or more bases or base pairs. Single and paired read degenerate mapping scores can be based on or derived from 100-mer alignability by GEM in ENCODE / CRG (Guigo), 75-mer alignability by GEM in ENCODE / CRG (Guigo), 100 base pair boxcar average of signal mappability, trajectory of paired read scores and maximum number of possible pairs, or a combination thereof.

[0079] Genomic region features may include one or more low average coverage regions from whole genome sequencing (WGS), zero average coverage regions from WGS, validated compressions, or combinations thereof. Low average coverage regions from WGS may include regions generated by Illumina® V3 chemistry, regions below the first percentile of a Poisson distribution based on average coverage, or combinations thereof. Zero average coverage regions from WGS may include regions generated by Illumina® V3 chemistry. Validated compressions may include regions with high mapping depth, regions with two or more observed haplotypes, regions predicted to be missing repeats in the reference, or combinations thereof. Genomic region features may include one or more alternative or non-reference sequences. The one or more alternative or non-reference sequences may include known structural variant junctions, known insertions, known deletions, alternative haplotypes, or combinations thereof. Genomic region features may include one or more gene phasing and rearranged genes. Examples of phasing and rearrangement genes include, but are not limited to, one or more major histocompatibility complex, blood group, and amylase gene families. The one or more major histocompatibility complexes may include one or more HLA class I, HLA class II, or a combination thereof. The one or more HLA class I may include HLA-A, HLA-B, HLA-C, or a combination thereof. The one or more HLA class II may include HLA-DP, HLA-DM, HLA-DOA, HLA-DOB, HLA-DQ, HLA-DR, or a combination thereof. The blood group genes may include ABO, RHD, RHCE, or a combination thereof.

[0080] Certain embodiments may include a nucleic acid sample or molecule comprising one or more genomic regions, at least one of which comprises a genomic region feature associated with the GC content of one or more nucleic acid molecules. GC content may refer to the GC content of a nucleic acid molecule. Alternatively, GC content may refer to the GC content of one or more nucleic acid molecules, and may be referred to as average GC content. As used herein, the terms "GC content" and "average GC content" may be used interchangeably. The GC content of a genomic region may be high GC content. Typically, a high GC content refers to a GC content of about 65%, 70%, 75%, 80%, 85%, 90%, 95%, 97%, or more. In some aspects of the present disclosure, a high GC content may refer to a GC content of about 70% or more. The GC content of a genomic region may be low GC content. Typically, low GC content refers to a GC content of about 65%, 60%, 55%, 50%, 45%, 40%, 35%, 30%, 25%, 20%, 15%, 10%, 5%, 2% or less.

[0081] Certain embodiments may include a nucleic acid sample or molecule comprising one or more genomic regions, at least one of which comprises a genomic region feature associated with the complexity of one or more nucleic acid molecules. The complexity of a nucleic acid molecule may refer to the randomness of the nucleotide sequence. Low complexity may refer to patterns, repetitions, and / or lack of one or more nucleotide species within the sequence.

[0082] Certain embodiments may include nucleic acid samples or molecules comprising one or more genomic regions, at least one of which comprises genomic region features related to the mappability of one or more nucleic acid molecules. The mappability of a nucleic acid molecule may refer to the specificity of its alignment to a reference sequence. A nucleic acid molecule with low mappability may be poorly aligned with a reference sequence.

[0083] D. Calculating the BAF distribution For each genomic region in the set of genomic regions, the B allele frequency (BAF) can be determined. BAF refers to a normalized measure of the allele intensity ratio of two alleles (A and B). A BAF of 100 or 0 indicates the complete absence of one of the two alleles (e.g., AA or BB). For example, if both chromosomes of a sample contain a nucleotide sequence variant relative to the reference genome, the BAF of a particular genomic region may be 100. Therefore, the variant allele must be present in nearly 100% of the reads for that sample. In some cases, the BAF distribution is normalized to generate a density plot such that the sum of the areas under the curve equals 1. Specifically, a histogram of the BAFs of all variants is created, the area of this histogram is calculated, and each bin of the histogram is divided by the calculated area.

[0084] For heterozygous alleles in a sample, the variant is present in a subset of sequence reads. Thus, a BAF of 50 indicates that both alleles (e.g., A, B) are equally present. In normal samples without copy number variation, a BAF of 0, 50, or 100 is expected. However, in tumor samples, mutations can alter BAF values. For example, a copy number event that duplicates the B allele may result in a B allele frequency of -67%. In another example, if only one chromosome contains a variant, the variant allele should be present in approximately 50% of the reads for that sample. Random variation inherent in the data generation process introduces noise, causing the observed BAF measurements to deviate from the ideal value for a particular copy number in both tumor and normal samples.

[0085] In some cases, a subset of genomic regions corresponding to heterozygous sites in a biological sample (e.g., an exome sample) is identified, and the allele frequency of each subset of genomic regions is quantified. Therefore, the VCF file can be filtered to obtain the reference and alternative read depths for all heterozygous sites with nucleotide sequence variants. Using this information, the BAF for each genomic region in the subset of genomic regions with heterozygosity can be calculated.

[0086] Based on the BAF determined for each set of genomic regions, a normalized BAF distribution can be calculated. The BAF distribution can indicate absolute amounts, percentages, and / or normalized amounts corresponding to the BAF of each genomic region in the set of genomic regions. In some cases, the normalized BAF distribution is determined from BAFs corresponding to a subset of genomic regions (e.g., heterozygous sites across the exome). Because BAF values typically range from 0 to 100, the normalized BAF distribution can include up to 101 BAF values, each representing a corresponding normalized frequency. Additionally or alternatively, the normalized BAF distribution can be modified so that corresponding BAF values can be binned into a predetermined number of bins. For example, each bin can correspond to a non-overlapping range of BAF values (e.g., 1 to 9, 10 to 19), and BAF values can be assigned to the corresponding bin. The BAF distribution can then be determined based on the values corresponding to each predetermined bin.

[0087] 7A-7E provide examples of plotted BAF distributions according to some embodiments. In both Figures 7A and 7B, the x-axis represents BAF values ranging from 0% (homozygous for the A allele) to 100% (homozygous for the B allele). As mentioned above, a BAF of 50 indicates heterozygosity, which corresponds to equal presence of both alleles (e.g., AB). In normal samples, BAFs of 0, 50, or 100 may be expected. However, in tumor samples, mutations may alter BAF values. For example, a copy number event that duplicates the B allele may result in a B allele frequency of -67%. The y-axis represents normalized frequency values corresponding to each BAF value on the x-axis. Normalized frequency values may identify several genomic regions corresponding to specific BAF values. For example, a normalized frequency of 50% BAF may be greater than 4.

[0088] Figure 7A shows a histogram of B allele frequencies from heterozygous sites in chromosomes of pure normal samples. The histogram shows a normal distribution, with most observed B allele frequencies falling near 50%. Figure 7B shows a histogram of B allele frequencies from heterozygous sites in pure tumor history samples. The histogram shows a shifted distribution, with more observed B allele frequencies falling further below 50%. As shown in Figure 7B, it can be seen that an increase in the number of somatic mutations may have altered the B allele frequency from 50%.

[0089] Figure 7C is a heat map, with each row representing the BAF distribution from the chromosomes of a pure normal sample. The heat map shows a normal distribution, with most observed B allele frequencies dropping near 50%. Figure 7E is a heat map, with each row representing the BAF distribution from the chromosomes of a pure tumor sample. The heat map shows an altered distribution, with more observed B allele frequencies dropping further below 50%. Figure 7D is a heat map from a sample that is 50% normal and 50% tumor. The heat map shows an intermediate distribution between the distributions shown in Figures 7C and 7E.

[0090] E. Classification of pure tumor and pure normal samples using BAF In some cases, the BAF signature of a biological sample is used to classify the biological sample as normal or tumorous. For example, a normalized BAF distribution can be calculated at heterozygous sites across the exomes of multiple pure tumor samples and pure normal samples. Pure tumor samples and pure normal samples can be classified using logistic regression using whole-exome BAF signatures.

[0091] 8 shows a plot 800 of the first two main components (PC1 and PC2) for classifying biological samples using B allele frequency features, according to some embodiments. In FIG. 8, the first two main components show near-linear separability between tumor and normal samples, indicating that BAF features may be suitable for estimating the tumor content of samples. Furthermore, it is demonstrated that pure tumor and pure normal samples can be classified using logistic regression using whole-exome BAF features.

[0092] F. Estimation of tumor purity based on BAF distribution A trained machine learning model may be used to process the BAF distribution of the biological sample to estimate a metric that identifies the tumor purity of the biological sample. In some embodiments, the trained machine learning model comprises a fully connected neural network. The fully connected neural network may include a fully connected layer with a rectified linear unit (ReLU) activation function. In some embodiments, the output activation function of the fully connected neural network is a sigmoid function. The loss function of the fully connected neural network may be configured to generate a mean squared error (MSE). In some embodiments, the fully connected neural network is tuned by hyperparameter search using random sampling with a linear search of layers, a linear search of size, a logarithmic search of learning rate, or a combination thereof.

[0093] In some cases, the trained machine learning model includes a one-dimensional convolutional neural network that can use the B allele frequency distribution as an input and can be configured to encode an input size of 25 height, 1 width, and 100 depth. In some embodiments, each layer of the one-dimensional convolutional neural network performs a 1×1 convolution followed by a ReLU activation function.

[0094] In some embodiments, the trained machine learning model includes a two-dimensional convolutional neural network. In some embodiments, the B allele frequency distribution of the two-dimensional convolutional neural network is used as an input and encoded into an input size defined as height 25, width 100, and depth 1. In some embodiments, each convolutional layer of the two-dimensional convolutional neural network is followed by a ReLU layer. The output of the two-dimensional convolutional neural network may be a densely connected layer using a sigmoid activation function. Furthermore, hyperparameters of the two-dimensional convolutional neural network may be tuned by adjusting the number of layers, filter size, number of filters, or a combination thereof.

[0095] An estimated metric identifying tumor purity may be output. For example, a report including the estimated metric may be output. In some embodiments, the report includes information identifying a B allele frequency distribution. The report may also include information identifying at least one diagnostic marker and / or at least one prognostic marker. In some embodiments, the report includes information identifying predicted somatic variants. The report may also include a treatment recommendation. In some embodiments, the treatment recommendation includes a recommendation to administer a treatment to the human subject. The treatment recommendation may include a recommendation not to administer a treatment to the human subject.

[0096] IV. Example Process for Estimating Tumor Purity from a Single Sample FIG. 9 includes a flowchart 900 illustrating an example of a method for estimating tumor purity of a biological sample according to certain embodiments. The operations described in flowchart 900 may be performed by a computer system implementing a trained machine learning model, such as a trained one-dimensional or two-dimensional convolutional neural network. While flowchart 900 may describe the operations as a sequential process, in various embodiments, many of the operations may be performed in parallel or simultaneously. The order of the operations may also be changed. The operations may include additional steps not shown. Furthermore, embodiments of this method may be implemented by hardware, software, firmware, middleware, microcode, hardware description languages, or any combination thereof. When implemented in software, firmware, middleware, or microcode, the program code or code segments for performing the associated tasks may be stored on a computer-readable medium, such as a storage medium.

[0097] At operation 910, the computer system obtains nucleic acid sequence data representing a plurality of nucleic acid molecules of a biological sample of a subject. The nucleic acid sequence data may be generated by sequencing a plurality of nucleic acid molecules of a tumor sample. In some cases, the plurality of nucleic acid molecules is isolated prior to sequencing. The nucleic acid sequence data may correspond to whole exome sequence data. Alternatively, or additionally, the nucleic acid sequence data is whole genome sequencing data.

[0098] In operation 920, the computer system aligns the nucleic acid sequence data to a reference genome. For example, a FASTQ file corresponding to the nucleic acid sequence data can be aligned to a reference genome to generate one or more BAM files.

[0099] At operation 930, the computer system identifies a set of genomic regions based on the aligned nucleic acid sequence data. Each genomic region of the set of genomic regions may contain one or more nucleotide sequence variants relative to a corresponding genomic region of the reference genome. Optionally, the computer system identifies one or more candidate nucleotide sequence variants in the nucleic acid sequence data and calculates a reference read depth and an alternative read depth for each of the one or more candidate nucleotide sequence variants.

[0100] In operation 940, the computer system measures the BAF of each genomic region of the set of genomic regions. BAF refers to a normalized measure of the allele intensity ratio of two alleles (A and B). In some cases, a BAF of 100 or 0 indicates the complete absence of one of the two alleles (e.g., AA or BB), while a BAF of 50 indicates equal presence of both alleles (e.g., AB). In normal samples, a BAF of 0, 50, or 100 may be expected. However, in tumor samples, mutations may alter BAF values.

[0101] In operation 950, the computer system determines a BAF distribution for the biological sample based on the BAFs of the set of genomic regions. Optionally, B allele frequencies are normalized.

[0102] At operation 960, the computer system processes the B allele frequency distribution using a trained machine learning model to estimate a metric identifying tumor purity of the biological sample. Optionally, the trained machine learning model was trained with a training dataset generated from nucleic acid sequence data from one or more tumor cells diluted with normal cells. The mean absolute error of the trained machine learning model may be less than about 0.2.

[0103] At operation 970, the computer system outputs metrics. In some cases, the computer system outputs a report including an estimated metric identifying tumor purity. In addition to the estimated metric, the report may include a measured B allele frequency distribution. The report may also include other types of information, including, but not limited to, at least one biomarker, at least one prognostic marker, predicted somatic variants, and a treatment recommendation (e.g., a recommendation as to whether a treatment should be administered to the subject).

[0104] V. Additional Considerations A. Probing Techniques Certain embodiments may include one or more labels. The one or more labels may be attached to one or more capture probes, nucleic acid molecules, beads, primers, or combinations thereof. Examples of labels include, but are not limited to, detectable labels such as radioisotopes, fluorescent dye molecules (fluorophores), chemiluminescent substances (chemiluminophores), chromophores, lumiphores, enzymes, colloidal particles, fluorescent microparticles, quantum dots, and one or more members of antigens, antibodies, haptens, avidin / streptavidin, biotin, haptens, enzyme cofactors / substrates, one or more members of a quenching system, chromogens, haptens, magnetic particles, materials exhibiting nonlinear optics, semiconductor nanocrystals, metal nanoparticles, enzymes, aptamers, and one or more members of a binding pair.

[0105] Certain embodiments may include one or more capture probes, multiple capture probes, or one or more capture probe sets. Typically, the capture probe comprises a nucleic acid binding site. The capture probe may further comprise one or more linkers. The capture probe may further comprise one or more labels. The one or more linkers may attach one or more labels to the nucleic acid binding site.

[0106] The capture probe may hybridize to one or more nucleic acid molecules in a sample. The capture probe may hybridize to one or more genomic regions. The capture probe may hybridize to one or more genomic regions within, surrounding, adjacent to, or spanning one or more genes, exons, introns, UTRs, or combinations thereof. The capture probe may hybridize to one or more genomic regions spanning one or more genes, exons, introns, UTRs, or combinations thereof. The capture probe may hybridize to one or more known indels. The capture probe may hybridize to one or more known structural variants.

[0107] Certain embodiments may comprise 1 or more, 2 or more, 3 or more, 4 or more, 5 or more, 6 or more, 7 or more, 8 or more, 9 or more, 10 or more, 20 or more, 30 or more, 40 or more, 50 or more, 60 or more, 70 or more, 80 or more, 90 or more, 100 or more, 125 or more, 150 or more, 175 or more, 200 or more, 250 or more, 300 or more, 350 or more, 400 or more, 500 or more, 600 or more, 700 or more, 800 or more, 900 or more, 1000 or more capture probes or capture probe sets, which may be different, similar, identical, or a combination thereof.

[0108] The one or more capture probes may comprise a nucleic acid binding site that hybridizes to at least a portion of one or more nucleic acid molecules or variants or derivatives thereof in a sample or subset of nucleic acid molecules. The capture probes may comprise a nucleic acid binding site that hybridizes to one or more genomic regions. The capture probes may hybridize to different, similar, and / or identical genomic regions. The one or more capture probes may be at least about 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 95%, 97%, 99% or more complementary to the one or more nucleic acid molecules or variants or derivatives thereof.

[0109] The capture probe can comprise one or more nucleotides. The capture probe can comprise 1 or more, 2 or more, 3 or more, 4 or more, 5 or more, 6 or more, 7 or more, 8 or more, 9 or more, 10 or more, 20 or more, 30 or more, 40 or more, 50 or more, 60 or more, 70 or more, 80 or more, 90 or more, 100 or more, 125 or more, 150 or more, 175 or more, 200 or more, 250 or more, 300 or more, 350 or more, 400 or more, 500 or more, 600 or more, 700 or more, 800 or more, 900 or more, or 1000 or more nucleotides. The capture probe can comprise about 100 nucleotides. The capture probe can comprise about 10 to about 500 nucleotides, about 20 to about 450 nucleotides, about 30 to about 400 nucleotides, about 40 to about 350 nucleotides, about 50 to about 300 nucleotides, about 60 to about 250 nucleotides, about 70 to about 200 nucleotides, or about 80 to about 150 nucleotides. In some embodiments of the present disclosure, the capture probe comprises about 80 nucleotides to about 100 nucleotides.

[0110] The plurality of capture probes or capture probe set may include two or more capture probes with identical, similar, and / or different nucleic acid binding site sequences, linkers, and / or labels. For example, two or more capture probes include the same nucleic acid binding site. In another example, two or more capture probes include similar nucleic acid binding sites. In another example, two or more capture probes include different nucleic acid binding sites. The two or more capture probes may further include one or more linkers. The two or more capture probes may further include different linkers. The two or more capture probes may further include similar linkers. The two or more capture probes may further include the same linker. The two or more capture probes may further include one or more labels. The two or more capture probes may further include different labels. The two or more capture probes may further include similar labels. The two or more capture probes may further include the same label.

[0111] B. Assay and Amplification Techniques Certain embodiments may include performing one or more assays on a sample containing one or more nucleic acid molecules. Generating two or more subsets of nucleic acid molecules may include performing one or more assays. The assays may be performed on a subset of nucleic acid molecules from the sample. The assays may be performed on one or more nucleic acid molecules from the sample. The assays may be performed on at least a portion of the subset of nucleic acid molecules. The assays may include one or more techniques, reagents, capture probes, primers, labels, and / or components for detection, quantification, and / or analysis of one or more nucleic acid molecules.

[0112] Assays may include, but are not limited to, sequencing, amplification, hybridization, enrichment, isolation, elution, fragmentation, detection, quantification of one or more nucleic acid molecules. Assays may include methods for preparing one or more nucleic acid molecules.

[0113] Certain embodiments may include performing one or more amplification reactions on one or more nucleic acid molecules in a sample. The term "amplification" refers to any process that generates at least one copy of a nucleic acid molecule. The terms "amplicon" and "amplified nucleic acid molecule" refer to copies of a nucleic acid molecule and may be used interchangeably. The amplification reaction may include a PCR-based method, a non-PCR-based method, or a combination thereof. Examples of non-PCR-based methods include, but are not limited to, multiple displacement amplification (MDA), transcription-mediated amplification (TMA), nucleic acid sequence-based amplification (NASBA), strand displacement amplification (SDA), real-time SDA, rolling circle amplification, or circle-to-circle amplification. PCR-based methods include, but are not limited to, PCR, HD-PCR, next-generation PCR, digital RTA, or any combination thereof. Additional PCR methods include, but are not limited to, linear amplification, allele-specific PCR, Alu PCR, assembly PCR, asymmetric PCR, droplet PCR, emulsion PCR, helicase-dependent amplification (HDA), hot-start PCR, inverse PCR, linear-post-exponential (LATE) PCR, long PCR, multiplex PCR, nested PCR, hemi-nested PCR, quantitative PCR, RT-PCR, real-time PCR, single-cell PCR, and touchdown PCR.

[0114] Certain embodiments may include performing one or more hybridization reactions on one or more nucleic acid molecules in a sample. The hybridization reaction may include hybridization of one or more capture probes to one or more nucleic acid molecules in a sample or subset of nucleic acid molecules. The hybridization reaction may include hybridizing one or more capture probe sets to one or more nucleic acid molecules in a sample or subset of nucleic acid molecules. The hybridization reaction may include one or more hybridization arrays, multiplex hybridization reactions, hybridization chain reactions, isothermal hybridization reactions, nucleic acid hybridization reactions, or combinations thereof. The one or more hybridization arrays may include hybridization array genotyping, hybridization array proportional sensing, DNA hybridization arrays, macroarrays, microarrays, high-density oligonucleotide arrays, genomic hybridization arrays, comparative hybridization arrays, or combinations thereof. A hybridization reaction can include one or more capture probes, one or more beads, one or more labels, one or more subsets of nucleic acid molecules, one or more nucleic acid samples, one or more reagents, one or more wash buffers, one or more elution buffers, one or more hybridization buffers, one or more hybridization chambers, one or more incubators, one or more separators, or combinations thereof.

[0115] Certain embodiments may include performing one or more enrichment reactions on one or more nucleic acid molecules in a sample. The enrichment reaction may include contacting the sample with one or more beads or bead sets. The enrichment reaction may include differential amplification of two or more subsets of nucleic acid molecules based on characteristics of one or more genomic regions. For example, the enrichment reaction may include differential amplification of two or more subsets of nucleic acid molecules based on GC content. Alternatively or additionally, the enrichment reaction may include differential amplification of two or more subsets of nucleic acid molecules based on methylation status. The enrichment reaction may include one or more hybridization reactions. The enrichment reaction may include isolation and / or purification of one or more hybridized nucleic acid molecules, one or more bead-bound nucleic acid molecules, one or more free nucleic acid molecules (e.g., nucleic acid molecules without capture probes, bead-unbound nucleic acid molecules), one or more labeled nucleic acid molecules, one or more unlabeled nucleic acid molecules, one or more amplicons, one or more unamplified nucleic acid molecules, or combinations thereof. Additionally or alternatively, the enrichment reaction may involve enrichment of one or more cell types in the sample. The one or more cell types may be enriched by flow cytometry.

[0116] One or more enrichment reactions can produce one or more enriched nucleic acid molecules. The enriched nucleic acid molecules can include nucleic acid molecules or variants or derivatives thereof. For example, the enriched nucleic acid molecules can include one or more hybridized nucleic acid molecules, one or more bead-bound nucleic acid molecules, one or more free nucleic acid molecules (e.g., nucleic acid molecules without capture probes, bead-unbound nucleic acid molecules), one or more labeled nucleic acid molecules, one or more unlabeled nucleic acid molecules, one or more amplicons, one or more unamplified nucleic acid molecules, or a combination thereof. The enriched nucleic acid molecules can be distinguished from non-enriched nucleic acid molecules by GC content, molecular size, genomic region, genomic region characteristics, or a combination thereof. The enriched nucleic acid molecules can be derived from one or more assays, supernatant, eluate, or a combination thereof. The enriched nucleic acid molecules can differ from non-enriched nucleic acid molecules by average size, average GC content, genomic region, or a combination thereof.

[0117] Certain embodiments may include performing one or more isolation or purification reactions on one or more nucleic acid molecules in a sample. The isolation or purification reaction may include contacting the sample with one or more beads or bead sets. The isolation or purification reaction may include one or more hybridization reactions, enrichment reactions, amplification reactions, sequencing reactions, or combinations thereof. The isolation or purification reaction may include the use of one or more separators. The one or more separators may include magnetic separators. The isolation or purification reaction may include separating nucleic acid molecules bound to the beads from nucleic acid molecules not bound to the beads. The isolation or purification reaction may include separating nucleic acid molecules hybridized with a capture probe from nucleic acid molecules not hybridized with a capture probe. The isolation or purification reaction may include separating a first subset of nucleic acid molecules from a second subset of nucleic acid molecules, the first subset of nucleic acid molecules differing from the second subset of nucleic acid molecules in average size, average GC content, genomic region, or a combination thereof.

[0118] Certain embodiments may include performing one or more elution reactions on one or more nucleic acid molecules in a sample. The elution reaction may include contacting the sample with one or more beads or bead sets. The elution reaction may include separating nucleic acid molecules bound to the beads from nucleic acid molecules not bound to the beads. The elution reaction may include separating nucleic acid molecules hybridized with a capture probe from nucleic acid molecules not hybridized with a capture probe. The elution reaction may include separating a first subset of nucleic acid molecules from a second subset of nucleic acid molecules, wherein the first subset of nucleic acid molecules differs from the second subset of nucleic acid molecules in average size, average GC content, genomic region, or a combination thereof.

[0119] Certain embodiments may include one or more fragmentation reactions. The fragmentation reaction may include fragmenting one or more nucleic acid molecules in a sample or subset of nucleic acid molecules to generate one or more fragmented nucleic acid molecules. The one or more nucleic acid molecules may be fragmented by sonication, needle shearing, nebulization, shearing (e.g., acoustic shearing, mechanical shearing, point-sink shearing), passage through a French pressure cell, or enzymatic digestion. Enzymatic digestion may occur by nuclease digestion (e.g., micrococcal nuclease digestion, endonuclease, exonuclease, RNAse H, or DNase I). The fragmentation of the one or more nucleic acid molecules may result in fragment sizes of about 100 base pairs to about 2000 base pairs, about 200 base pairs to about 1500 base pairs, about 200 base pairs to about 1000 base pairs, about 200 base pairs to about 500 base pairs, about 500 base pairs to about 1500 base pairs, and about 500 base pairs to about 1000 base pairs. The one or more fragmentation reactions may result in fragments of about 50 base pairs to about 1000 base pairs in size. The one or more fragmentation reactions may result in fragment sizes of about 100 base pairs, 150 base pairs, 200 base pairs, 250 base pairs, 300 base pairs, 350 base pairs, 400 base pairs, 450 base pairs, 500 base pairs, 550 base pairs, 600 base pairs, 650 base pairs, 700 base pairs, 750 base pairs, 800 base pairs, 850 base pairs, 900 base pairs, 950 base pairs, 1000 base pairs or more.

[0120] Fragmenting the one or more nucleic acid molecules can include mechanically shearing one or more nucleic acid molecules in a sample for a period of time, which can be at least about 10 seconds, 15 seconds, 20 seconds, 25 seconds, 30 seconds, 35 seconds, 40 seconds, 45 seconds, 50 seconds, 55 seconds, 60 seconds, 65 seconds, 70 seconds, 75 seconds, 80 seconds, 85 seconds, 90 seconds, 95 seconds, 100 seconds, 125 seconds, 150 seconds, 175 seconds, 200 seconds, 225 seconds, 250 seconds, 275 seconds, 300 seconds, 325 seconds, 350 seconds, 375 seconds, 400 seconds, 425 seconds, 450 seconds, 475 seconds, 500 seconds, or more.

[0121] Fragmenting the one or more nucleic acid molecules can include contacting the nucleic acid sample with one or more beads. Fragmenting the one or more nucleic acid molecules can include contacting the nucleic acid sample with a plurality of beads, wherein the ratio of the volume of the plurality of beads to the volume of the nucleic acid sample is about 0.10, 0.20, 0.30, 0.40, 0.50, 0.60, 0.70, 0.80, 0.90, 1.00, 1.10, 1.20, 1.30, 1.40, 1.50, 1.60, 1.70, 1.80, 1.90, 2.00 or more. Fragmenting the one or more nucleic acid molecules can include contacting the nucleic acid sample with a plurality of beads, wherein the ratio of the volume of the plurality of beads to the volume of the nucleic acid sample is about 2.00, 1.90, 1.80, 1.70, 1.60, 1.50, 1.40, 1.30, 1.20, 1.10, 1.00, 0.90, 0.80, 0.70, 0.60, 0.50, 0.40, 0.30, 0.20, 0.10, 0.05, 0.04, 0.03, 0.02, 0.01 or less.

[0122] Certain embodiments may include performing one or more detection reactions on one or more nucleic acid molecules in a sample. The detection reaction may include one or more sequencing reactions. Alternatively, performing the detection reaction includes optical sensing, electrical sensing, or a combination thereof. Optical sensing may include optical sensing of photoluminescence photon emission, fluorescence photon emission, pyrophosphate photon emission, chemiluminescence photon emission, or a combination thereof. Electrical sensing may include electrical sensing of ion concentration, ion current modulation, nucleotide electric field, nucleotide tunneling current, or a combination thereof.

[0123] Certain embodiments may include performing one or more quantification reactions on one or more nucleic acid molecules in a sample. The quantification reactions may include sequencing, PCR, qPCR, digital PCR, or a combination thereof.

[0124] Certain embodiments may include one or more samples. Certain embodiments may include 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100 or more samples. The samples may be derived from a subject. The two or more samples may be derived from a single subject. The two or more samples may be derived from 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100 or more different subjects. The subject may be a mammal, reptile, amphibian, bird, or fish. Mammals may be humans, apes, orangutans, monkeys, chimpanzees, cows, pigs, horses, rodents, birds, reptiles, dogs, cats, or other animals. Reptiles may be lizards, snakes, alligator snapping turtles, turtles, crocodiles, tortoises, etc. Amphibians may be toads, frogs, newts, and salamanders. Examples of birds include, but are not limited to, ducks, geese, penguins, ostriches, and owls. Examples of fish include, but are not limited to, catfish, eels, sharks, and swordfish. Preferably, the subject is a human. The subject may be suffering from a disease or condition (e.g., cancer).

[0125] The two or more samples may be taken at 1 hour, 2 hours, 3 hours, 4 hours, 5 hours, 6 hours, 7 hours, 8 hours, 9 hours, 10 hours, 11 hours, 12 hours, 15 hours, 20 hours, 30 hours, 40 hours, 50 hours, 60 hours, 70 hours, 80 hours, 90 hours, 100 hours, 200 hours, 300 hours, 400 hours, 500 hours, 600 hours, 700 hours, 800 hours, 900 hours, 1000 hours or more. The time points may range from 1 hour, 2 hours, 3 hours, 4 hours, 5 hours, 6 hours, 7 hours, 8 hours, 9 hours, 10 hours, 11 hours, 12 hours, 13 hours, 14 hours, 15 hours, 16 hours, 17 hours, 18 hours, 19 hours, 20 hours, 21 hours, 22 hours, 23 hours, 24 hours, 25 hours, 30 hours, 35 hours, 40 hours, 45 hours, 50 hours, 55 hours, 60 hours or more. The time points may span 1 day, 2 days, 3 days, 4 days, 5 days, 6 days, 7 days, 8 days, 9 days, 10 days, 11 days, 12 days, 13 days, 14 days, 15 days, 16 days, 17 days, 18 days, 19 days, 20 days, 21 days, 22 days, 23 days, 24 days, 25 days, 30 days, 35 days, 40 days, 45 days, 50 days, 55 days, 60 days or more. The time points may range from 1 week, 2 weeks, 3 weeks, 4 weeks, 5 weeks, 6 weeks, 7 weeks, 8 weeks, 9 weeks, 10 weeks, 11 weeks, 12 weeks, 13 weeks, 14 weeks, 15 weeks, 16 weeks, 17 weeks, 18 weeks, 19 weeks, 20 weeks, 21 weeks, 22 weeks, 23 weeks, 24 weeks, 25 weeks, 30 weeks, 35 weeks, 40 weeks, 45 weeks, 50 weeks, 55 weeks, 60 weeks or more. The time points may range from 1 month, 2 months, 3 months, 4 months, 5 months, 6 months, 7 months, 8 months, 9 months, 10 months, 11 months, 12 months, 13 months, 14 months, 15 months, 16 months, 17 months, 18 months, 19 months, 20 months, 21 months, 22 months, 23 months, 24 months, 25 months, 30 months, 35 months, 40 months, 45 months, 50 months, 55 months, 60 months or more.The time points may span 1 year, 2 years, 3 years, 4 years, 5 years, 6 years, 7 years, 8 years, 9 years, 10 years, 11 years, 12 years, 13 years, 14 years, 15 years, 16 years, 17 years, 18 years, 19 years, 20 years, 21 years, 22 years, 23 years, 24 years, 25 years, 30 years, 35 years, 40 years, 45 years, 50 years, 55 years, 60 years or more.

[0126] The sample may be from a bodily fluid, cell, skin, tissue, organ, or a combination thereof. The sample may be blood, plasma, a blood fraction, saliva, sputum, urine, semen, vaginal fluid, cerebrospinal fluid, feces, cell, or tissue biopsy. The sample may be from the adrenal gland, adnexa, bladder, brain, ear, esophagus, eye, gallbladder, heart, kidney, large intestine, liver, lung, oral cavity, muscle, nasal cavity, pancreas, parathyroid gland, pineal gland, pituitary gland, skin, small intestine, spleen, stomach, thymus, thyroid gland, trachea, uterus, appendix, cornea, skin, heart valve, artery, or vein.

[0127] The sample may contain one or more nucleic acid molecules. The nucleic acid molecules may be DNA molecules, RNA molecules (e.g., mRNA, cRNA, or miRNA), and DNA / RNA hybrids. Examples of DNA molecules include, but are not limited to, double-stranded DNA, single-stranded DNA, single-stranded DNA hairpins, cDNA, and genomic DNA. The nucleic acids may be RNA molecules such as double-stranded RNA, single-stranded RNA, ncRNA, RNA hairpins, and mRNA. Examples of ncRNA include, but are not limited to, siRNA, miRNA, snoRNA, piRNA, tiRNA, PASR, TASR, aTASR, TSSa-RNA, snRNA, RE-RNA, uaRNA, x-ncRNA, hY RNA, usRNA, snaR, and vtRNA.

[0128] Certain embodiments may include one or more containers. Certain embodiments may include 1 or more, 2 or more, 3 or more, 4 or more, 5 or more, 6 or more, 7 or more, 8 or more, 9 or more, 10 or more, 20 or more, 30 or more, 40 or more, 50 or more, 60 or more, 70 or more, 80 or more, 90 or more, 100 or more, 125 or more, 150 or more, 175 or more, 200 or more, 250 or more, 300 or more, 350 or more, 400 or more, 500 or more, 600 or more, 700 or more, 800 or more, 900 or more, or 1000 or more containers. The one or more containers may be different, similar, identical, or a combination thereof. Examples of containers include, but are not limited to, plates, microplates, PCR plates, wells, microwells, tubes, Eppendorf tubes, vials, arrays, microarrays, and chips.

[0129] Certain embodiments may include one or more reagents. Certain embodiments may include 1 or more, 2 or more, 3 or more, 4 or more, 5 or more, 6 or more, 7 or more, 8 or more, 9 or more, 10 or more, 20 or more, 30 or more, 40 or more, 50 or more, 60 or more, 70 or more, 80 or more, 90 or more, 100 or more, 125 or more, 150 or more, 175 or more, 200 or more, 250 or more, 300 or more, 350 or more, 400 or more, 500 or more, 600 or more, 700 or more, 800 or more, 900 or more, or 1000 or more reagents. The one or more reagents may be different, similar, identical, or a combination thereof. The reagent may improve the efficiency of one or more assays. The reagent may improve the stability of the nucleic acid molecule or variant or derivative thereof. Reagents may include, but are not limited to, enzymes, proteases, nucleases, molecules, polymerases, reverse transcriptases, ligases, and chemical compounds. Certain embodiments may include performing an assay that includes one or more antioxidants. Generally, an antioxidant is a molecule that inhibits the oxidation of another molecule. Examples of antioxidants include, but are not limited to, ascorbic acid (e.g., vitamin C), glutathione, lipoic acid, uric acid, carotenes, α-tocopherol (e.g., vitamin E), ubiquinol (e.g., coenzyme Q), and vitamin A.

[0130] Certain embodiments may include one or more buffers or solutions. The one or more buffers or solutions may be different, similar, identical, or a combination thereof. The buffers or solutions may improve the efficiency of one or more assays. The buffers or solutions may improve the stability of the nucleic acid molecule or variants or derivatives thereof. Buffers or solutions may include, but are not limited to, wash buffers, elution buffers, and hybridization buffers.

[0131] Certain embodiments may include one or more beads, multiple beads, or one or more bead sets. Certain embodiments may include one or more, two or more, three or more, four or more, five or more, six or more, seven or more, eight or more, nine or more, ten or more, twenty or more, thirty or more, forty or more, fifty or more, sixty or more, seventy or more, eighty or more, ninety or more, one hundred or more, one or more beads or bead sets. The one or more beads or bead sets may be different, similar, identical, or a combination thereof. The beads may be magnetic, antibody-coated, protein A-crosslinked, protein G-crosslinked, streptavidin-coated, oligonucleotide-linked, silica-coated, or a combination thereof. Examples of beads include, but are not limited to, AMPure® beads, AMPure® XP beads, streptavidin beads, agarose beads, magnetic beads, Dynabeads®, MACS® microbeads, antibody-conjugated beads (e.g., anti-immunoglobulin microbeads), protein A-conjugated beads, protein G-conjugated beads, protein A / G-conjugated beads, protein L-conjugated beads, oligo dT-conjugated beads, silica beads, silica-like beads, anti-biotin microbeads, anti-fluorescent chromium microbeads, and BcMag™ carboxy-terminated magnetic beads. In some embodiments of the present disclosure, the one or more beads comprise one or more AMPure® beads. Alternatively or additionally, the one or more beads comprise AMPure® XP beads.

[0132] Certain embodiments may include one or more primers, multiple primers, or one or more primer sets. Primers may further include one or more linkers. Primers may further include one or more labels. Primers may be used in one or more assays. For example, primers are used in one or more sequencing reactions, amplification reactions, or combinations thereof. Certain embodiments may include one or more, 2 or more, 3 or more, 4 or more, 5 or more, 6 or more, 7 or more, 8 or more, 9 or more, 10 or more, 20 or more, 30 or more, 40 or more, 50 or more, 60 or more, 70 or more, 80 or more, 90 or more, 100 or more, 125 or more, 150 or more, 175 or more, 200 or more, 250 or more, 300 or more, 350 or more, 400 or more, 500 or more, 600 or more, 700 or more, 800 or more, 900 or more, 1000 or more primers or primer sets. The primers may comprise approximately 100 nucleotides. The primers may comprise about 10 to about 500 nucleotides, about 20 to about 450 nucleotides, about 30 to about 400 nucleotides, about 40 to about 350 nucleotides, about 50 to about 300 nucleotides, about 60 to about 250 nucleotides, about 70 to about 200 nucleotides, or about 80 to about 150 nucleotides. In some embodiments of the present disclosure, the primers comprise about 80 to about 100 nucleotides. The one or more primers or primer sets may be different, similar, identical, or a combination thereof.

[0133] The primers hybridize to at least a portion of one or more nucleic acid molecules or variants or derivatives thereof in a sample or subset of nucleic acid molecules. The primers may hybridize to one or more genomic regions. The primers may hybridize to different, similar, and / or identical genomic regions. The one or more primers may be at least about 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 95%, 97%, 99% or more complementary to the one or more nucleic acid molecules or variants or derivatives thereof.

[0134] The primer may comprise one or more nucleotides. The primer may comprise 1 or more, 2 or more, 3 or more, 4 or more, 5 or more, 6 or more, 7 or more, 8 or more, 9 or more, 10 or more, 20 or more, 30 or more, 40 or more, 50 or more, 60 or more, 70 or more, 80 or more, 90 or more, 100 or more, 125 or more, 150 or more, 175 or more, 200 or more, 250 or more, 300 or more, 350 or more, 400 or more, 500 or more, 600 or more, 700 or more, 800 or more, 900 or more, or 1000 or more nucleotides. The primer may comprise about 100 nucleotides. The primer may comprise about 10 to about 500 nucleotides, about 20 to about 450 nucleotides, about 30 to about 400 nucleotides, about 40 to about 350 nucleotides, about 50 to about 300 nucleotides, about 60 to about 250 nucleotides, about 70 to about 200 nucleotides, or about 80 to about 150 nucleotides. In some embodiments of the present disclosure, the primer comprises about 80 nucleotides to about 100 nucleotides.

[0135] The plurality of primers or primer set may include two or more primers having the same, similar, and / or different sequences, linkers, and / or labels. For example, two or more primers include the same sequence. In another example, two or more primers include similar sequences. In yet another example, two or more primers include different sequences. The two or more primers may further include one or more linkers. The two or more primers may further include different linkers. The two or more primers may further include similar linkers. The two or more primers may further include the same linkers. The two or more primers may further include one or more labels. The two or more primers may further include different labels. The two or more primers may further include similar labels. The two or more primers may further include the same labels.

[0136] The capture probe, primer, label, and / or bead may comprise one or more nucleotides that may include RNA, DNA, a mixture of DNA and RNA residues, or modified analogs such as 2'-0Me, 2'-fluoro (2'-F), locked nucleic acid (LNA), or abasic sites.

[0137] Certain embodiments may include one or more labels. Certain embodiments may include 1 or more, 2 or more, 3 or more, 4 or more, 5 or more, 6 or more, 7 or more, 8 or more, 9 or more, 10 or more, 20 or more, 30 or more, 40 or more, 50 or more, 60 or more, 70 or more, 80 or more, 90 or more, 100 or more, 125 or more, 150 or more, 175 or more, 200 or more, 250 or more, 300 or more, 350 or more, 400 or more, 500 or more, 600 or more, 700 or more, 800 or more, 900 or more, 1000 or more labels. The one or more labels may be different, similar, identical, or a combination thereof.

[0138] Examples of labels include, but are not limited to, chemical, biochemical, biological, colorimetric, enzymatic, fluorescent, and luminescent labels known in the art, including dyes, photocrosslinkers, cytotoxic compounds, drugs, affinity labels, photoaffinity labels, reactive compounds, antibodies or antibody fragments, biomaterials, nanoparticles, spin labels, fluorophores, metal-containing moieties, radioactive moieties, novel functional groups, groups that interact covalently or noncovalently with other molecules, photocaged moieties, actinically excitable moieties, ligands, photoisomerizable moieties, biotin, biotin analogs, moieties incorporating heavy atoms, chemically cleavable groups, photocleavable groups, redox-active agents, isotope-labeled moieties, biophysical probes, phosphorescent groups, chemiluminescent groups, electron-dense groups, magnetic groups, intercalating groups, chromophores, energy transfer agents, biologically active agents, detectable labels, or combinations thereof.

[0139] The label may be a chemical label. Examples of chemical labels may include, but are not limited to, biotin and radioisotopes (e.g., iodine, carbon, phosphate, hydrogen).

[0140] The methods, kits, and compositions disclosed herein can include biological labels, which can include metabolic labels, including, but not limited to, bioorthogonal azide-modified amino acids, sugars, and other compounds.

[0141] The methods, kits, and compositions disclosed herein may include an enzyme label, including, but not limited to, horseradish peroxidase (HRP), alkaline phosphatase (AP), glucose oxidase, and O-galactosidase. The enzyme label may be luciferase.

[0142] The methods, kits, and compositions disclosed herein may include a biological label. The fluorescent label may be an organic dye (e.g., FITC), a biological fluorophore (e.g., green fluorescent protein), or a quantum dot. A non-limiting list of fluorescent labels includes fluorescein isothiocyanate (FITC), DyLight® Fluor, fluorescein, rhodamine (tetramethylrhodamine isothiocyanate, TRITC), coumarin, Lucifer Yellow, and BODIPY. The label may also be a fluorophore. Examples of fluorophores include indocarbocyanine (C3), indodicarbocyanine (C5), Cy3, Cy3.5, Cy5, Cy5.5, Cy7, Texas Red, Pacific Blue, Oregon Green 488, Alexa Fluor® 355, Alexa Fluor® 488, Alexa Fluor® 532, Alexa Fluor® 546, Alexa Fluor® 555, Alexa Fluor® 568, Alexa Fluor® 594, Alexa Fluor® 647, Alexa Fluor® 660, Alexa Fluor® 594 ... Fluorescent labels include, but are not limited to, Fluor® 680, JOE, Lissamine, rhodamine green, BODIPY, fluorescein isothiocyanate (FITC), carboxyfluorescein (FAM), phycoerythrin, rhodamine, dichlororhodamine (dRhodamine), carboxytetramethylrhodamine (TAMRA), carboxy-X-rhodamine (ROX™), LIZ™, VIC™, NED™, PET™, SYBR, PicoGreen®, RiboGreen®, etc. The fluorescent label can be green fluorescent protein (GFP), red fluorescent protein (RFP), yellow fluorescent protein, phycobiliproteins (e.g., allophycocyanin, phycocyanin, phycoerythrin, and phycoerythrocyanin).

[0143] Certain embodiments may include one or more linkers. Certain embodiments may include 1 or more, 2 or more, 3 or more, 4 or more, 5 or more, 6 or more, 7 or more, 8 or more, 9 or more, 10 or more, 20 or more, 30 or more, 40 or more, 50 or more, 60 or more, 70 or more, 80 or more, 90 or more, 100 or more, 125 or more, 150 or more, 175 or more, 200 or more, 250 or more, 300 or more, 350 or more, 400 or more, 500 or more, 600 or more, 700 or more, 800 or more, 900 or more, 1000 or more linkers. The one or more linkers may be different, similar, identical, or a combination thereof.

[0144] Suitable linkers include any chemical or biological compound that can be attached to a label, primer, and / or capture probe disclosed herein. When the linker is attached to both the label and the primer or capture probe, the linker allows sufficient separation between the label and the primer or capture probe. A suitable linker does not significantly interfere with the ability of the primer and / or capture probe to hybridize to a nucleic acid molecule, a portion thereof, or a variant or derivative thereof. A suitable linker does not significantly interfere with the ability of the label to be detected. The linker can be rigid. The linker can be flexible. The linker can be semi-rigid. The linker can be proteolytically stable (e.g., resistant to proteolytic cleavage). The linker can be proteolytically unstable (e.g., susceptible to proteolytic cleavage). The linker can be helical. The linker can be non-helical. The linker can be coiled. The linker can be triple-stranded. The linker can include a turn conformation. The linker may be a single chain. The linker may be a long chain. The linker may be a short chain. The linker may comprise at least about 5 residues, at least about 10 residues, at least about 15 residues, at least about 20 residues, at least about 25 residues, at least about 30 residues, or at least about 40 residues or more.

[0145] Examples of linkers include, but are not limited to, hydrazones, disulfides, thioethers, and peptide linkers. The linker may be a peptide linker. The peptide linker may include a proline residue. The peptide linker may include arginine, phenylalanine, threonine, glutamine, glutamic acid, or any combination thereof. The linker may be a heterobifunctional crosslinker.

[0146] Certain embodiments include performing 1 or more, 2 or more, 3 or more, 4 or more, 5 or more, 6 or more, 7 or more, 8 or more, 9 or more, 10 or more, 11 or more, 12 or more, 13 or more, 14 or more, 15 or more, 20 or more, 25 or more, 30 or more, 35 or more, 40 or more, 45 or more, or 50 or more assays on a sample containing one or more nucleic acid molecules. The two or more assays can be different, similar, identical, or a combination thereof. For example, certain embodiments include performing two or more sequencing reactions. In another example, certain embodiments include performing two or more assays, where at least one of the two or more assays comprises a sequencing reaction. In yet another example, certain embodiments include performing two or more assays, where at least two of the two or more assays comprise a sequencing reaction and a hybridization reaction. The two or more assays may be performed sequentially, simultaneously, or a combination thereof. For example, the two or more sequencing reactions may be performed simultaneously. In another example, certain embodiments include performing a hybridization reaction followed by a sequencing reaction. In yet another example, certain embodiments include simultaneously performing two or more hybridization reactions followed by two or more sequencing reactions. The two or more assays may be performed by one or more devices. For example, two or more amplification reactions may be performed by a PCR machine. In another example, two or more sequencing reactions may be performed by two or more sequencers.

[0147] C. Device Certain embodiments may include one or more devices. Certain embodiments may include one or more assays including one or more devices. Certain embodiments may include the use of one or more devices to perform one or more steps or assays. Certain embodiments may include the use of one or more devices in one or more steps or assays. For example, performing a sequencing reaction may include one or more sequencers. In another example, generating a subset of nucleic acid molecules may include the use of one or more magnetic separators. In yet another example, one or more processors may be used in analyzing one or more nucleic acid samples. Examples of devices include, but are not limited to, sequencers, thermocyclers, real-time PCR instruments, magnetic separators, transmission devices, hybridization chambers, electrophoresis devices, centrifuges, microscopes, imagers, fluorometers, luminometers, plate readers, computers, processors, and bioanalyzers.

[0148] Certain embodiments may include one or more sequencers. The one or more sequencers may include one or more HiSeq, MiSeq, HiScan, Genome Analyzer IIx, SOLiD Sequencer, Ion Torrent PGM, 454 GS Junior, PacBio RS, or a combination thereof. The one or more sequencers may include one or more sequencing platforms. The one or more sequencing platforms may include GS FLX 454 by Life Technologies / Roche, Genome Analyzer by Solexa / Illumina, SOLiD by Applied Biosystems, CGA Platform by Complete Genomics, PacBio RS by Pacific Biosciences, or a combination thereof.

[0149] Certain embodiments may include one or more thermocyclers. The one or more thermocyclers may be used for amplifying one or more nucleic acid molecules. Certain embodiments may include one or more real-time PCR instruments. The one or more real-time PCR instruments may include a thermal cycler and a fluorometer. The one or more thermocyclers may be used for amplifying and detecting one or more nucleic acid molecules.

[0150] Certain embodiments may include one or more magnetic separators that may be used to separate paramagnetic and ferromagnetic particles from the suspension. The one or more magnetic separators may include one or more LifeStep™ Biomagnetic Separators, SPHERO™ FlexiMag Separators, SPHERO™ MicroMag Separators, SPHERO™ HandiMag Separators, SPHERO™ MiniTube Mag Separators, SPHERO™ UltraMag Separators, DynaMag™ Magnets, DynaMag™-2 Magnets, or combinations thereof.

[0151] Certain embodiments may include one or more bioanalyzers. Generally, a bioanalyzer is a chip-based capillary electrophoresis device capable of analyzing RNA, DNA, and proteins. The one or more bioanalyzers may include an Agilent 2100 bioanalyzer.

[0152] Certain embodiments may include one or more processors. The one or more processors may analyze, compile, store, sort, combine, evaluate, or otherwise process one or more data and / or results from one or more assays, one or more data and / or results based on or derived from one or more assays, one or more outputs from one or more assays, one or more outputs based on or derived from one or more assays, one or more outputs from one or more data and / or results, one or more outputs based on or derived from one or more data and / or results, or combinations thereof. The one or more processors may transmit one or more data, results, or outputs from one or more assays, one or more data, results, or outputs based on or derived from one or more assays, one or more outputs from one or more data or results, one or more outputs based on or derived from one or more data or results, or combinations thereof. The one or more processors may receive and / or store requests from a user. The one or more processors may create or generate one or more data, results, or outputs. The one or more processors may create or generate one or more biomedical reports. The one or more processors may transmit one or more biomedical reports. The one or more processors may analyze, compile, store, sort, combine, evaluate, or otherwise process information from one or more databases, one or more data or results, one or more outputs, or a combination thereof. The one or more processors may analyze, compile, store, sort, combine, evaluate, or otherwise process information from 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 30, or more databases. The one or more processors may transmit one or more requests, data, results, outputs, and / or information to one or more users, processors, computers, computer systems, memory locations, devices, databases, or combinations thereof.The one or more processors may receive one or more requests, data, results, output, and / or information from one or more users, processors, computers, computer systems, memory locations, devices, databases, or combinations thereof. The one or more processors may retrieve one or more requests, data, results, output, and / or information from one or more users, processors, computers, computer systems, memory locations, devices, databases, or combinations thereof.

[0153] Certain embodiments may include one or more memory locations that may store information, data, results, output, requests, or combinations thereof. The one or more memory locations may receive information, data, results, output, requests, or combinations thereof from one or more users, processors, computers, computer systems, devices, or combinations thereof.

[0154] The methods described herein may be implemented using one or more computers and / or computer systems. A computer or computer system may include electronic storage locations (e.g., databases, memory) with machine-executable code for implementing the methods provided herein, and one or more processors for executing the machine-executable code.

[0155] The code may be pre-compiled and configured for use on a device with a processor adapted to execute the code, or it may be compiled at run time. The code may be provided in a selectable programming language so that the code can be executed in pre-compiled or compiled form.

[0156] The one or more computers and / or computer systems may analyze, compile, store, sort, combine, evaluate, or otherwise process one or more data and / or results from one or more assays, one or more data and / or results based on or derived from one or more assays, one or more outputs from one or more assays, one or more outputs based on or derived from one or more assays, one or more outputs from one or more data and / or results, one or more outputs based on or derived from one or more data and / or results, or combinations thereof. The one or more computers and / or computer systems may transmit one or more data, results, or outputs from one or more assays, one or more data, results, or outputs based on or derived from one or more assays, one or more outputs from one or more data or results, one or more outputs based on or derived from one or more data or results, or combinations thereof. The one or more computers and / or computer systems may receive and / or store requests from a user. The one or more computers and / or computer systems may create or generate one or more data, results, or outputs. The one or more computers and / or computer systems may create or generate one or more biomedical reports. The one or more computers and / or computer systems may transmit one or more biomedical reports. The one or more computers and / or computer systems may analyze, compile, store, sort, combine, evaluate, or otherwise process information from one or more databases, one or more data or results, one or more outputs, or a combination thereof. The one or more computers and / or computer systems may analyze, compile, store, sort, combine, evaluate, or otherwise process information from 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 30, or more databases.The one or more computers and / or computer systems may transmit one or more requests, data, results, output, and / or information to one or more users, processors, computers, computer systems, memory locations, devices, or combinations thereof. The one or more computers and / or computer systems may receive one or more requests, data, results, output, and / or information from one or more users, processors, computers, computer systems, memory locations, devices, or combinations thereof. The one or more computers and / or computer systems may retrieve one or more requests, data, results, output, and / or information from one or more users, processors, computers, computer systems, memory locations, devices, databases, or combinations thereof.

[0157] D. Database Certain embodiments may include one or more databases. Certain embodiments may include at least about 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 30, or more databases. The databases may include genomic databases, proteomic databases, pharmacogenomic databases, biomedical databases, and scientific databases. The databases may be public databases. Alternatively, or additionally, the databases may include proprietary databases. The databases may be commercial databases. Such databases include, but are not limited to, Cosmic, GnomAD, Dbsnp, Mills Indels, MendelDB, PharmGKB, Varimed, Regulome, BreakSeq (curated junctions), OMIM (Online Mendelian Inheritance in Man), HGMD (Human Genome Mutation Database), NCBI db SNP, NCBI RefSeq, GENCODE, GO (Gene Ontology), and KEGG (Kyoto Encyclopedia of Genes and Genomes).

[0158] Certain embodiments may include analyzing one or more databases. Certain embodiments may include analyzing at least about 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 30, or more databases. Analyzing one or more databases may include one or more algorithms, computers, processors, memory locations, devices, or combinations thereof.

[0159] Certain embodiments may include identifying one or more nucleic acid regions based on data and / or information from one or more databases. Certain embodiments may include identifying one or more sets of nucleic acid regions based on data and / or information from one or more databases. Certain embodiments may include identifying one or more nucleic acid regions and / or sets of nucleic acid regions based on data and / or information from at least about two or more databases. Certain embodiments may include identifying one or more nucleic acid regions and / or sets of nucleic acid regions based on data and / or information from at least about three or more databases. Certain embodiments may include identifying one or more nucleic acid regions and / or sets of nucleic acid regions based on data and / or information from at least about 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 30, or more databases.

[0160] Certain embodiments may include analyzing one or more results based on data and / or information from one or more databases. Certain embodiments may include analyzing one or more sets of results based on data and / or information from one or more databases. Certain embodiments may include analyzing one or more combined results based on data and / or information from one or more databases. Certain embodiments may include analyzing one or more results, sets of results, or combined results based on data and / or information from at least about two or more databases. Certain embodiments may include analyzing one or more results, sets of results, or combined results based on data and / or information from at least about three or more databases. Certain embodiments may include analyzing one or more results, sets of results, or combined results based on data and / or information from at least about 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 30, or more databases.

[0161] Certain embodiments may include comparing one or more results based on data and / or information from one or more databases. Certain embodiments may include comparing one or more sets of results based on data and / or information from one or more databases. Certain embodiments may include comparing one or more combined results based on data and / or information from one or more databases. Certain embodiments may include comparing one or more results, sets of results, or combined results based on data and / or information from at least about two or more databases. Certain embodiments may include comparing one or more results, sets of results, or combined results based on data and / or information from at least about three or more databases. Certain embodiments may include comparing one or more results, sets of results, or combined results based on data and / or information from at least about 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 30, or more databases.

[0162] Certain embodiments may include biomedical databases, genomic databases, biomedical reports, disease reports, case-control analyses, and rare variant detection analyses based on data and / or information from one or more databases, one or more assays, one or more data or results, one or more outputs based on or derived from one or more assays, one or more outputs based on or derived from one or more data or results, or combinations thereof.

[0163] E. Datasets and Analysis Certain embodiments may include one or more data, one or more datasets, one or more combined data, one or more combined datasets, one or more results, one or more result sets, one or more combined results, or a combination thereof. The data and / or results may be based on or derived from one or more assays, one or more databases, or a combination thereof. Certain embodiments may include analysis of one or more data, one or more datasets, one or more combined data, one or more combined datasets, one or more results, one or more result sets, one or more combined results, or a combination thereof. Certain embodiments may include processing of one or more data, one or more datasets, one or more combined data, one or more combined datasets, one or more results, one or more result sets, one or more combined results, or a combination thereof.

[0164] Certain embodiments may include at least one analysis and at least one processing of one or more data, one or more datasets, one or more combined data, one or more combined datasets, one or more results, one or more result sets, one or more combined results, or a combination thereof. Certain embodiments may include one or more analyses and one or more processing of one or more data, one or more datasets, one or more combined data, one or more combined datasets, one or more results, one or more result sets, one or more combined results, or a combination thereof. Certain embodiments may include at least 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 30, 40, 50, 60, 70, 80, 90, 100, 200, 300, 400, 500, 600, 700, 800, 900, 1000 or more different analyses of one or more data, one or more datasets, one or more combined data, one or more combined datasets, one or more results, one or more sets of results, one or more combined results, or combinations thereof. Certain embodiments may include at least 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 30, 40, 50, 60, 70, 80, 90, 100, 200, 300, 400, 500, 600, 700, 800, 900, 1000 or more different processing of one or more data, one or more datasets, one or more combined data, one or more combined datasets, one or more results, one or more sets of results, one or more combined results, or combinations thereof, wherein the one or more analyses and / or one or more processing may be performed simultaneously, sequentially, or a combination thereof.

[0165] The one or more analyses and / or one or more treatments may be performed at 1 hour, 2 hours, 3 hours, 4 hours, 5 hours, 6 hours, 7 hours, 8 hours, 9 hours, 10 hours, 11 hours, 12 hours, 15 hours, 20 hours, 30 hours, 40 hours, 50 hours, 60 hours, 70 hours, 80 hours, 90 hours, 100 hours, 200 hours, 300 hours, 400 hours, 500 hours, 600 hours, 700 hours, 800 hours, 900 hours, 1000 hours or more. The time points may range from 1 hour, 2 hours, 3 hours, 4 hours, 5 hours, 6 hours, 7 hours, 8 hours, 9 hours, 10 hours, 11 hours, 12 hours, 13 hours, 14 hours, 15 hours, 16 hours, 17 hours, 18 hours, 19 hours, 20 hours, 21 hours, 22 hours, 23 hours, 24 hours, 25 hours, 30 hours, 35 hours, 40 hours, 45 hours, 50 hours, 55 hours, 60 hours or more. The time points may span 1 day, 2 days, 3 days, 4 days, 5 days, 6 days, 7 days, 8 days, 9 days, 10 days, 11 days, 12 days, 13 days, 14 days, 15 days, 16 days, 17 days, 18 days, 19 days, 20 days, 21 days, 22 days, 23 days, 24 days, 25 days, 30 days, 35 days, 40 days, 45 days, 50 days, 55 days, 60 days or more. The time points may range from 1 week, 2 weeks, 3 weeks, 4 weeks, 5 weeks, 6 weeks, 7 weeks, 8 weeks, 9 weeks, 10 weeks, 11 weeks, 12 weeks, 13 weeks, 14 weeks, 15 weeks, 16 weeks, 17 weeks, 18 weeks, 19 weeks, 20 weeks, 21 weeks, 22 weeks, 23 weeks, 24 weeks, 25 weeks, 30 weeks, 35 weeks, 40 weeks, 45 weeks, 50 weeks, 55 weeks, 60 weeks or more. The time points may range from 1 month, 2 months, 3 months, 4 months, 5 months, 6 months, 7 months, 8 months, 9 months, 10 months, 11 months, 12 months, 13 months, 14 months, 15 months, 16 months, 17 months, 18 months, 19 months, 20 months, 21 months, 22 months, 23 months, 24 months, 25 months, 30 months, 35 months, 40 months, 45 months, 50 months, 55 months, 60 months or more.The time points may span 1 year, 2 years, 3 years, 4 years, 5 years, 6 years, 7 years, 8 years, 9 years, 10 years, 11 years, 12 years, 13 years, 14 years, 15 years, 16 years, 17 years, 18 years, 19 years, 20 years, 21 years, 22 years, 23 years, 24 years, 25 years, 30 years, 35 years, 40 years, 45 years, 50 years, 55 years, 60 years or more.

[0166] Certain embodiments may include one or more data. The one or more data may include one or more raw data based on or derived from one or more assays. The one or more data may include one or more raw data based on or derived from one or more databases. The one or more data may include at least partially analyzed data based on or derived from one or more raw data. The one or more data may include at least partially processed data based on or derived from one or more raw data. The one or more data may include fully analyzed data based on or derived from one or more raw data. The one or more data may include fully processed data based on or derived from one or more raw data. The data may include sequencing read data or expression data. The data may include biomedical, chemical, pharmacological, and / or genetic information.

[0167] Certain embodiments may include one or more combined data. The one or more combined data may include two or more data sets. The one or more combined data may include one or more raw data based on or derived from one or more assays. The one or more combined data may include one or more raw data based on or derived from one or more databases. The one or more combined data may include at least partially analyzed data based on or derived from one or more raw data. The one or more combined data may include at least partially processed data based on or derived from one or more raw data. The one or more combined data may include fully analyzed data based on or derived from one or more raw data. The one or more combined data may include fully processed data based on or derived from one or more raw data. The one or more combined data may include sequencing read data or expression data. The one or more combined data may include biomedical, chemical, pharmacological, and / or genetic information.

[0168] Certain embodiments may include one or more datasets. The one or more datasets may include one or more data. The one or more datasets may include one or more combined data. The one or more datasets may include one or more raw data based on or derived from one or more assays. The one or more datasets may include one or more raw data based on or derived from one or more databases. The one or more datasets may include at least partially analyzed data based on or derived from one or more raw data. The one or more datasets may include at least partially processed data based on or derived from one or more raw data. The one or more datasets may include fully analyzed data based on or derived from one or more raw data. The one or more datasets may include fully processed data based on or derived from one or more raw data. The datasets may include sequencing read data or expression data. The datasets may include biomedical, chemical, pharmacological, and / or genetic information.

[0169] Certain embodiments may include one or more combined datasets. The one or more combined datasets may include two or more data. The one or more combined datasets may include two or more combined data. The one or more combined datasets may include two or more datasets. The one or more combined datasets may include one or more raw data based on or derived from one or more assays. The one or more combined datasets may include one or more raw data based on or derived from one or more databases. The one or more combined datasets may include at least partially analyzed data based on or derived from one or more raw data. The one or more combined datasets may include at least partially processed data based on or derived from one or more raw data. The one or more combined datasets may include fully analyzed data based on or derived from one or more raw data. The one or more combined datasets may include fully processed data based on or derived from one or more raw data. Certain embodiments may further include further processing and / or analysis of the combined datasets. The one or more combined datasets may include sequencing read data or expression data. The one or more combined datasets may include biomedical, chemical, pharmacological, and / or genetic information.

[0170] Certain embodiments may include one or more results. The one or more results may include one or more data, datasets, combined data, and / or combined datasets. The one or more results may be based on or derived from one or more data, datasets, combined data, and / or combined datasets. The one or more results may be generated from one or more assays. The one or more results may be based on or derived from one or more assays. The one or more results may be based on or derived from one or more databases. The one or more results may include at least partially analyzed results based on or derived from one or more data, datasets, combined data, and / or combined datasets. The one or more results may include at least partially processed results based on or derived from one or more data, datasets, combined data, and / or combined datasets. The one or more results may include fully analyzed results based on or derived from one or more data, datasets, combined data, and / or combined datasets. The one or more results may include one or more data, datasets, combined data, and / or fully processed results based on or derived from combined datasets. The results may include sequencing read data or expression data. The results may include biomedical, chemical, pharmacological, and / or genetic information.

[0171] Certain embodiments may include one or more result sets. The one or more result sets may include one or more data, datasets, combined data, and / or combined datasets. The one or more result sets may be based on or derived from one or more data, datasets, combined data, and / or combined datasets. The one or more result sets may be generated from one or more assays. The one or more result sets may be based on or derived from one or more assays. The one or more result sets may be based on or derived from one or more databases. The one or more result sets may include at least partially analyzed result sets based on or derived from one or more data, datasets, combined data, and / or combined datasets. The one or more result sets may include at least partially processed result sets based on or derived from one or more data, datasets, combined data, and / or combined datasets. The one or more result sets may include fully analyzed result sets based on or derived from one or more data, datasets, combined data, and / or combined datasets. The one or more result sets may include one or more data, datasets, combined data, and / or fully processed result sets based on or derived from combined datasets. The result sets may include sequencing read data or expression data. The result sets may include biomedical, chemical, pharmacological, and / or genetic information.

[0172] Certain embodiments may include one or more combined results. The combined results may include one or more results, sets of results, and / or sets of combined results. The combined results may be based on or derived from one or more results, sets of results, and / or sets of combined results. The one or more combined results may include one or more data, datasets, combined data, and / or combined datasets. The one or more combined results may be based on or derived from one or more data, datasets, combined data, and / or combined datasets. The one or more combined results may be generated from one or more assays. The one or more combined results may be based on or derived from one or more assays. The one or more combined results may be based on or derived from one or more databases. The one or more combined results may include at least partially analyzed combined results based on or derived from one or more data, datasets, combined data, and / or combined datasets. The one or more combined results may include at least partially processed combined results based on or derived from one or more data, datasets, combined data, and / or combined datasets. The one or more combined results may include fully analyzed combined results based on or derived from one or more data, datasets, combined data, and / or combined datasets. The one or more combined results may include fully processed combined results based on or derived from one or more data, datasets, combined data, and / or combined datasets. The combined results may include sequencing read data or expression data. The combined results may include biomedical, chemical, pharmacological, and / or genetic information.

[0173] Certain embodiments may include one or more combined result sets. The combined result set may include one or more results, result sets, and / or combined results. The combined result set may be based on or derived from one or more results, result sets, and / or combined results. The one or more combined result sets may include one or more data, datasets, combined data, and / or combined datasets. The one or more combined result sets may be based on or derived from one or more data, datasets, combined data, and / or combined datasets. The one or more combined result sets may be generated from one or more assays. The one or more combined result sets may be based on or derived from one or more assays. The one or more combined result sets may be based on or derived from one or more databases. The one or more combined result sets may include at least partially analyzed combined result sets based on or derived from one or more data, datasets, combined data, and / or combined datasets. The one or more combined result sets may include at least partially processed combined result sets based on or derived from one or more data, datasets, combined data, and / or combined datasets. The one or more combined result sets may include fully analyzed combined result sets based on or derived from one or more data, datasets, combined data, and / or combined datasets. The one or more combined result sets may include fully processed combined result sets based on or derived from one or more data, datasets, combined data, and / or combined datasets. The combined result sets may include sequencing read data or expression data. The combined result sets may include biomedical, chemical, pharmacological, and / or genetic information.

[0174] Certain embodiments may include one or more outputs, sets of outputs, combined outputs, and / or combined sets of outputs. Methods, libraries, kits, and systems described herein may include generating one or more outputs, sets of outputs, combined outputs, and / or combined sets of outputs. The set of outputs may include one or more outputs, one or more combined outputs, or a combination thereof. The combined output may include one or more outputs, one or more sets of outputs, one or more combined sets of outputs, or a combination thereof. The combined set of outputs may include one or more outputs, one or more sets of outputs, one or more combined outputs, or a combination thereof. The one or more outputs, sets of outputs, combined outputs, and / or combined sets of outputs may be based on or derived from one or more data, one or more datasets, one or more combined data, one or more combined datasets, one or more results, one or more sets of results, one or more combined results, or a combination thereof. The one or more outputs, sets of outputs, combined outputs, and / or sets of combined outputs may be based on or derived from one or more databases, and the one or more outputs, sets of outputs, combined outputs, and / or sets of combined outputs may include one or more biomedical reports, biomedical outputs, rare variant outputs, pharmacogenetic outputs, population study outputs, case-control outputs, biomedical databases, genomic databases, disease databases, net content.

[0175] Certain embodiments may include one or more biomedical outputs, one or more sets of biomedical outputs, one or more combined biomedical outputs, or one or more sets of combined biomedical outputs. The methods, libraries, kits, and systems described herein may include generating one or more biomedical outputs, one or more sets of biomedical outputs, one or more combined biomedical outputs, or one or more sets of combined biomedical outputs. The set of biomedical outputs may include one or more biomedical outputs, one or more combined biomedical outputs, or a combination thereof. The combined biomedical outputs may include one or more biomedical outputs, one or more sets of biomedical outputs, one or more combined biomedical outputs, or a combination thereof. The combined set of biomedical outputs may include one or more biomedical outputs, one or more sets of biomedical outputs, one or more combined biomedical outputs, or a combination thereof. The one or more biomedical outputs, one or more sets of biomedical outputs, one or more combined biomedical outputs, or one or more sets of combined biomedical outputs may be based on or derived from one or more data, one or more datasets, one or more combined data, one or more combined datasets, one or more results, one or more sets of results, one or more combined results, one or more outputs, one or more sets of outputs, one or more combined outputs, one or more sets of combined outputs, or combinations thereof. The one or more biomedical outputs may include subject biomedical information. The subject biomedical information may predict, diagnose, and / or prognose one or more biomedical characteristics. The one or more biomedical characteristics may include severity of a disease or condition, genetic risk of a disease or condition, reproductive risk, genetic risk to a fetus, risk of adverse drug reactions, efficacy of drug therapy, prediction of optimal drug dosage, transplant tolerance, or a combination thereof.

[0176] Certain embodiments may include one or more biomedical reports. The methods, libraries, kits, and systems described herein may include generating one or more biomedical reports. The one or more biomedical reports may be based on or derived from one or more data, one or more datasets, one or more combined data, one or more combined datasets, one or more results, one or more sets of results, one or more combined results, one or more outputs, one or more sets of outputs, one or more combined outputs, one or more sets of combined outputs, one or more biomedical outputs, one or more sets of biomedical outputs, combined biomedical outputs, one or more sets of biomedical outputs, or combinations thereof. The biomedical report may predict, diagnose, and / or prognose one or more biomedical characteristics. The one or more biomedical characteristics may include severity of a disease or condition, genetic risk of a disease or condition, reproductive risk, genetic risk to a fetus, risk of adverse drug reactions, efficacy of drug therapy, prediction of optimal drug dosage, transplant tolerance, or a combination thereof.

[0177] Certain embodiments may also include transmitting one or more data, information, results, outputs, reports, or combinations thereof. For example, data / information based on or derived from one or more assays may be transmitted to another device and / or instrument. In another example, the data, results, outputs, biomedical outputs, biomedical reports, or combinations thereof may be transmitted to another device and / or instrument. Information derived from an algorithm may also be transmitted to another device and / or instrument. Information based on analysis of one or more databases may be transmitted to another device and / or instrument. The transmission of the data / information may include transferring data / information from a first source to a second source. The first source and the second source may be in the same approximate location (e.g., in the same room, building, block, campus). Alternatively, the first source and the second source may be in multiple locations (e.g., multiple cities, states, countries, continents, etc.). The data, results, outputs, biomedical outputs, biomedical reports may be communicated to a patient and / or a healthcare professional.

[0178] The transmission may be based on an analysis of one or more data, results, information, databases, outputs, reports, or a combination thereof. For example, the transmission of a second report is based on an analysis of a first report. Alternatively, the transmission of a report is based on an analysis of one or more data or results. The transmission may also be based on receiving one or more requests. For example, the transmission of a report may be based on receiving a request from a user (e.g., a patient, a healthcare professional, an individual).

[0179] Data / information transmission may include digital or analog transmission. Digital transmission may involve the physical transmission of data (digital bit streams) over point-to-point or point-to-multipoint communication channels. Examples of such channels include copper wire, fiber optics, wireless communication channels, and storage media. Data may be represented as electromagnetic signals such as electrical voltages, radio waves, microwaves, or infrared signals.

[0180] Analog transmission may involve the transmission of continuously varying analog signals. Messages may be represented either as a series of pulses in a line code (baseband transmission) or as a limited set of continuously varying waveforms using digital modulation schemes (passband transmission). Passband modulation and corresponding demodulation (also known as detection) may be performed by modem equipment. According to the most common definition of a digital signal, both baseband and passband signals representing bit streams are considered digital transmissions, while another definition considers only baseband signals digital and passband transmission of digital data a form of digital-to-analog conversion.

[0181] Certain embodiments may include one or more sample identifiers. The sample identifiers may include labels, barcodes, and other indicia that may be associated with one or more samples and / or subsets of nucleic acid molecules. Certain embodiments may include one or more processors, one or more memory locations, one or more computers, one or more monitors, one or more computer software, and one or more algorithms for associating data, results, output, biomedical output, and / or biomedical reports with the samples.

[0182] Certain embodiments may include a processor for correlating expression levels of one or more nucleic acid molecules with a prognosis of disease outcome. Certain embodiments may include one or more of a variety of correlation techniques, including lookup tables, algorithms, multivariate models, and linear or non-linear combinations of formula models or algorithms. The expression levels may be converted into one or more likelihood scores reflecting the likelihood that the patient providing the sample will exhibit a particular disease outcome. The models and / or algorithms may be provided in machine-readable format and may optionally further specify treatment for the patient or patient class.

[0183] In some cases, the methods and systems described herein are used to generate output including detection and / or quantification of genomic DNA regions, such as regions containing DNA polymorphisms (e.g., germline variants or somatic variants). In some cases, the detection of one or more genomic regions is based on one or more algorithms, depending on the source of the data input or database, as described elsewhere herein. Each of the one or more algorithms may be used to receive, combine, and generate data including the detection of genomic regions (i.e., polymorphisms). In some embodiments, the methods and systems of the present invention may include the detection of genomic regions based on one or more, two or more, three or more, four or more, five or more, six or more, seven or more, eight or more, nine or more, or ten or more algorithms. The algorithms may be machine learning algorithms, computer-implemented algorithms, machine-executed algorithms, automated algorithms, etc.

[0184] The resulting data for each nucleic acid sample can be analyzed using feature selection methods, including filter methods that assess feature relevance by examining inherent properties of the data, wrapper methods that embed model hypotheses within the feature subset search, and embedded methods where the search for the optimal feature set is built into an algorithm or model.

[0185] In some cases, the detection of one or more genomic regions is based on one or more statistical models. Statistical models or filtering techniques useful in the methods of the present invention include: (1) parametric methods such as two-sample t-tests, ANOVA analysis, Bayesian frameworks, and the use of gamma distribution models; (2) model-free methods such as Wilcoxon rank sum tests, inter- and intra-class sum-of-squares tests, rank product methods, random pure quantification methods, or the use of TNoM, which involves setting a threshold point for the difference in expression fold change between two datasets and then detecting the threshold point for each gene that minimizes the number of misclassifications; and (3) multivariate methods such as bivariate methods, correlation-based feature selection (CFS), minimum redundancy maximum relevance (MRMR), Markov blanket filters, Markov models, hidden Markov models (HMMs), and uncorrelated shrunken centroid (USC) methods. In some cases, the hidden Markov model (HMM) is given an internal state, and the internal state is set according to the overall copy number of chromosomes in the first or second nucleic acid sample. In one example, for a diploid chromosome, the internal states of the HMM can be homozygous deletion (locally zero copies), heterozygous deletion (locally one copy), normal (locally two copies), duplication (two or more copies), and reference gap (present as a state to distinguish gaps from homozygous deletions). In another example, for a haploid chromosome (e.g., a male X or Y), the internal states of the HMM can be homozygous deletion (locally zero copies), normal (locally two copies), duplication (two or more copies), and reference gap (present as a state to distinguish gaps from homozygous deletions). For example, for a haploid chromosome, no heterozygous deletion state may be available. In another example, for trisomy and / or tetrasomy, additional intermediate HMM states may have additional intermediate states, which may account for various CNV possibilities. In another embodiment, a hidden Markov model is used to filter the output by examining the insertion size of reads measured near the breakpoints of the detected feature.

[0186] Other models or algorithms useful in the methods of the present invention include sequential search methods, genetic algorithms, estimation of distribution algorithms, random forest algorithms, weight vectors in support vector machine algorithms, and weights in logistic regression algorithms. Bioinformatics. 2007 Oct 1;23(19):2507-17 provides an overview of the relative merits of the algorithms or models provided above for data analysis. Illustrative algorithms include, but are not limited to, methods that reduce the number of variables, such as principal component analysis algorithms; methods that directly handle a large number of variables, such as partial least squares, independent component analysis algorithms, and statistical methods; and methods based on machine learning techniques. Statistical methods include penalized logistic regression, predictive analysis of microarrays (PAM), methods based on shrunken centroids, support vector machine analysis, and regularized linear discriminant analysis.

[0187] In some embodiments, HMM-based detection algorithms can detect large or substantially large CNVs "segmentally." In some cases, fluctuations in coverage signal can result in small detection gaps along the length of the true CNV. In one example, a 1 megabase pair (Mbp) deletion can be detected as a few separate nominal detections with a small gap between them. To mitigate this, a merging operation can be used to identify pairs of adjacent detections separated by a gap smaller than either of the two bracketed detections. The merging operation then measures the median coverage level within the gap. If the median coverage exceeds a predefined threshold, the two detections are merged into a single large detection spanning the two original detections (including the enclosed detection gap). In one example, the true feature spans both detections, and the gap is a statistical artifact. Using actual sequencing data from a sample known to have a large CNV, this merging operation can significantly increase fidelity regarding the actual characteristics of the CNV.

[0188] The methods and systems provided herein may further include the use of feature selection algorithms as provided herein. In some embodiments of the invention, feature selection is performed using the LIMMA software package (Smyth, GK (2005). Limma: linear models for microarray data. In: Bioinformatics and Computational Biology Solutions using R and Bioconductor, R. Gentleman, V. Carey, S. Dudoit, R. Irizarry, W. Huber (eds.), Springer, New York, pages 397-420).

[0189] In some embodiments of the present invention, diagonal linear discriminant analysis, K-nearest neighbor algorithm, support vector machine (SVM) algorithm, linear support vector machine, random forest algorithm, probabilistic model-based method, or combinations thereof are provided to detect one or more genomic regions. In some embodiments, identified markers that distinguish samples (e.g., disease vs. normal) or genomic regions (e.g., copy number variation vs. normal) are selected based on the statistical significance of the difference in expression levels between the classes of interest. In some cases, the statistical significance is adjusted by applying Benjamini-Hochberg or another correction for false discovery rate (FDR).

[0190] In some cases, the algorithms may be complemented with meta-analysis approaches such as those described in Fishel and Kaufman et al. 2007 Bioinformatics 23(13): 1599-606. In some cases, the algorithms may be complemented with meta-analysis approaches such as reproducibility analysis. In some cases, the reproducibility analysis selects markers that appear in at least one predicted expression product marker set.

[0191] Statistical evaluation of the detection of genomic regions provides one or more quantitative values indicating one or more of the following: likelihood of diagnostic accuracy; likelihood of injury, disease, condition, etc.; likelihood of a particular injury, disease, or condition; and likelihood of success of a particular therapeutic intervention. Thus, physicians, who likely lack training in genetics or molecular biology, do not need to understand the raw data. Rather, the data are presented directly to physicians in the form of quantitative values to guide patient care. The results can be statistically evaluated using several methods known in the art, including, but not limited to, Student's t-test, two-tailed t-test, Pearson rank sum analysis, hidden Markov model analysis, QQ plot analysis, principal component analysis, one-way analysis of variance, two-way analysis of variance, and LIMMA.

[0192] F. Diseases and Conditions Certain embodiments may include predicting, diagnosing, and / or prognosing the extent or outcome of a disease or condition in a subject based on one or more biomedical outputs. Predicting, diagnosing, and / or prognosing the extent or outcome of a disease in a subject may include diagnosing a disease or condition, identifying a disease or condition, determining the stage of a disease or condition, assessing the risk of a disease or condition, assessing the risk of disease recurrence, assessing drug efficacy, assessing the risk of adverse drug reactions, predicting optimal drug dosage, predicting drug resistance, or a combination thereof.

[0193] The sample disclosed herein may be from a subject suffering from cancer. The sample may include malignant tissue, benign tissue, or a mixture thereof. The cancer may be recurrent and / or refractory cancer. Examples of cancer include, but are not limited to, sarcoma, carcinoma, lymphoma, or leukemia. In some cases, a sample containing cancerous tissue is obtained, but a matching normal sample is not available. In some cases, no matching normal sample is available. In some cases, a matching normal sample is obtained (e.g., for the purpose of training and testing the models disclosed herein).

[0194] Sarcomas are cancers of bone, cartilage, fat, muscle, blood vessels, or other connective or supportive tissues. Sarcomas include, but are not limited to, bone cancer, fibrosarcoma, chondrosarcoma, Ewing's sarcoma, malignant hemangioendothelioma, malignant schwannoma, bilateral vestibular schwannoma, osteosarcoma, and soft tissue sarcomas (e.g., alveolar soft part sarcoma, angiosarcoma, cystosarcoma phyllodes, dermatofibrosarcoma, desmoid tumor, epithelioid sarcoma, extraskeletal osteosarcoma, fibrosarcoma, hemangiopericytoma, angiosarcoma, Kaposi's sarcoma, leiomyosarcoma, liposarcoma, lymphangiosarcoma, lymphosarcoma, malignant fibrous histiocytoma, neurofibrosarcoma, rhabdomyosarcoma, and synovial sarcoma).

[0195] Carcinomas are cancers that arise from epithelial cells, which are cells that cover the surface of the body, produce hormones, and make up glands. Non-limiting examples of carcinomas include breast cancer, pancreatic cancer, lung cancer, colon cancer, colorectal cancer, rectal cancer, kidney cancer, bladder cancer, stomach cancer, prostate cancer, liver cancer, ovarian cancer, brain cancer, vaginal cancer, vulvar cancer, uterine cancer, oral cancer, penile cancer, testicular cancer, esophageal cancer, skin cancer, fallopian tube cancer, head and neck cancer, gastrointestinal stromal cancer, adenocarcinoma, cutaneous or intraocular melanoma, anal region cancer, small intestine cancer, endocrine system cancer, thyroid cancer, parathyroid cancer, adrenal gland cancer, urethral cancer, renal pelvis cancer, ureter cancer, endometrial cancer, cervical cancer, pituitary cancer, neoplasms of the central nervous system (CNS), primary CNS lymphoma, brain stem glioma, and spinal axis tumor. The cancer can be a skin cancer such as basal cell carcinoma, squamous cell carcinoma, melanoma, non-melanoma, or actinic (solar) keratosis.

[0196] The cancer may be lung cancer. Lung cancer may occur in the airways that branch off from the trachea and supply the lungs (bronchi) or the small air sacs (alveoli) of the lungs. Lung cancer includes non-small cell lung cancer (NSCLC), small cell lung cancer, and mesothelioma. Examples of NSCLC include squamous cell carcinoma, adenocarcinoma, and large cell carcinoma. Mesothelioma may be a cancerous tumor of the lining of the lung and chest cavity (pleura) or the abdominal lining (peritoneum). Mesothelioma may be caused by asbestos exposure. The cancer may be a brain tumor, such as glioblastoma.

[0197] The cancer may be a central nervous system (CNS) tumor. CNS tumors may be classified as gliomas or non-gliomas. Gliomas may be malignant gliomas, high-grade gliomas, or diffuse intrinsic pontine gliomas. Examples of gliomas include astrocytomas, oligodendrogliomas (or a mixture of oligodendrogliomas and astrocytomas), and ependymomas. Astrocytomas include, but are not limited to, low-grade astrocytomas, anaplastic astrocytomas, glioblastoma multiforme, pilocytic astrocytomas, pleomorphic xanthoastrocytomas, and subependymal giant cell astrocytomas. Oligodendrogliomas include low-grade oligodendrogliomas (or oligodendrogliomas) and anaplastic oligodendrogliomas. Non-gliomas include meningiomas, pituitary adenomas, primary CNS lymphomas, and medulloblastomas. The cancer may be a meningioma.

[0198] The leukemia can be acute lymphocytic leukemia, acute myeloid leukemia, chronic lymphocytic leukemia, or chronic myelogenous leukemia. Additional types of leukemia include hairy cell leukemia, chronic myelomonocytic leukemia, and juvenile myelomonocytic leukemia.

[0199] Lymphoma is a cancer of lymphocytes and can arise from either B or T lymphocytes. The two main types of lymphoma are Hodgkin's lymphoma, formerly known as Hodgkin's disease, and non-Hodgkin's lymphoma. Hodgkin's lymphoma is characterized by the presence of Reed-Sternberg cells. Non-Hodgkin's lymphoma is any lymphoma that is not Hodgkin's lymphoma. Non-Hodgkin's lymphoma can be indolent or aggressive. Non-Hodgkin's lymphomas include, but are not limited to, diffuse large B-cell lymphoma, follicular lymphoma, mucosa-associated lymphoid tissue lymphoma (MALT), small cell lymphocytic lymphoma, mantle cell lymphoma, Burkitt's lymphoma, mediastinal large B-cell lymphoma, Waldenstrom's macroglobulinemia, nodal marginal zone B-cell lymphoma (NMZL), splenic marginal zone lymphoma (SMZL), extranodal marginal zone B-cell lymphoma, intravascular large B-cell lymphoma, primary effusion lymphoma, and lymphomatoid granulomatosis.

[0200] Certain embodiments may include treating and / or preventing a disease or condition in a subject based on one or more biomedical outputs. The one or more biomedical outputs may recommend one or more therapies. The one or more biomedical outputs may suggest, select, prescribe, recommend, or otherwise determine a course of treatment and / or prevention for a disease or condition. The one or more biomedical outputs may recommend modifying or continuing one or more therapies. Modifying one or more therapies may include administering, initiating, reducing, increasing, and / or terminating one or more therapies. The one or more therapies may include anti-cancer therapy, anti-viral therapy, anti-bacterial therapy, anti-fungal therapy, immunosuppressive therapy, or a combination thereof. The one or more therapies may treat, alleviate, or prevent one or more diseases or indications.

[0201] Examples of anti-cancer therapies include, but are not limited to, surgery, chemotherapy, radiation therapy, immunotherapy / biological therapy, and photodynamic therapy. Anti-cancer therapies may include chemotherapeutic agents, monoclonal antibodies (e.g., rituximab, trastuzumab), cancer vaccines (e.g., therapeutic vaccines, prophylactic vaccines), gene therapy, or a combination thereof.

[0202] G. Systems, Kits, and Libraries The methods of the present disclosure may be implemented by a system, a kit, a library, or a combination thereof. The methods of the present disclosure may include one or more systems. The systems of the present disclosure may be implemented via a kit, a library, or both. A system may include one or more components for implementing any of the methods or steps of a particular embodiment. For example, a system may include one or more kits, devices, libraries, or a combination thereof. A system may include one or more sequencers, processors, memory locations, computers, computer systems, or a combination thereof. A system may include a transmission device.

[0203] The kits may include various reagents for performing various operations disclosed herein, including sample processing and / or analytical operations. The kits may include instructions for performing at least some of the operations disclosed herein. The kits may include one or more capture probes, one or more beads, one or more labels, one or more linkers, one or more devices, one or more reagents, one or more buffers, one or more samples, one or more databases, or combinations thereof.

[0204] The library may include one or more capture probes. The library may include one or more subsets of nucleic acid molecules. The library may include one or more databases. The library may be created or generated from any of the methods, kits, or systems disclosed herein. A database library may be created from one or more databases. A method of creating one or more libraries may include (a) aggregating information from one or more databases to create an aggregated dataset, (b) analyzing the aggregated dataset, and (c) creating one or more database libraries from the aggregated dataset.

[0205] VI. Computing Environment 10 illustrates an example computer system 1000 for implementing some of the embodiments disclosed herein. The computer system 1000 may have a distributed architecture in which some components (e.g., memory and processor) are part of an end-user device and some other similar components (e.g., memory and processor) are part of a computer server. The computer system 1000 includes at least a processor 1002, a memory 1004, a storage device 1006, input / output (I / O) peripherals 1008, communication peripherals 1010, and an interface bus 1012. The interface bus 1012 is configured to communicate, transmit, and transfer data, control, and commands between the various components of the computer system 1000. The processor 1002 may include one or more processing units, such as a CPU, a GPU, a TPU, a systolic array, or a SIMD processor. The memory 1004 and storage device 1006 include computer-readable storage media such as RAM, ROM, electrically erasable programmable read-only memory (EEPROM), hard drives, CD-ROMs, optical storage devices, magnetic storage devices, electronic non-volatile computer storage devices (e.g., Flash®), and other tangible storage media. Any such computer-readable storage media may be configured to store instructions or program code embodying aspects of the present disclosure. The memory 1004 and storage device 1006 also include computer-readable signal media. Computer-readable signal media include a propagated data signal embodied with computer-readable program code. Such a propagated signal may take any of a variety of forms, including, but not limited to, electromagnetic, optical, or any combination thereof. Computer-readable signal media includes any computer-readable medium that is not a computer-readable storage medium and can communicate, propagate, or transmit a program for use in connection with the computer system 1000.

[0206] Additionally, memory 1004 includes an operating system, programs, and applications. Processor 1002 is configured to execute stored instructions and includes, for example, a logic processing unit, microprocessor, digital signal processor, and other processors. Memory 1004 and / or processor 1002 can be virtualized and / or hosted within another computing system, for example, in a cloud network or data center. I / O peripherals 1008 include user interfaces such as keyboards, screens (e.g., touchscreens), microphones, speakers, and other input / output devices, as well as computing components such as graphical processing units, serial ports, parallel ports, universal serial buses, and other input / output peripherals. I / O peripherals 1008 are connected to processor 1002 via any ports coupled to interface bus 1012. Communications peripherals 1010 are configured to facilitate communication between computer system 1000 and other computing devices over a communications network and include, for example, network interface controllers, modems, wireless and wired interface cards, antennas, and other communications peripherals.

[0207] While the subject matter of the present invention has been described in detail with reference to specific embodiments thereof, those skilled in the art will understand that, upon gaining the foregoing understanding, modifications, variations, and equivalents of such embodiments may readily occur to those skilled in the art. Accordingly, it should be understood that the present disclosure is presented for purposes of example and not limitation, and is not intended to exclude the inclusion of such modifications, variations, and / or additions to the subject matter of the present invention as would be readily apparent to one of ordinary skill in the art. Indeed, the methods and systems described herein may be embodied in a variety of other forms, and various omissions, substitutions, and changes in the form of the methods and systems described herein may be made without departing from the spirit of the disclosure. The accompanying claims and their equivalents are intended to cover such forms or modifications as are within the scope and spirit of the present disclosure.

[0208] Unless otherwise specified, use of terms such as "processing," "computing," "calculating," "determining," and "identifying" will be understood throughout the description herein to refer to the actions or processes of a computing device, such as one or more computers or similar electronic computing devices, that manipulate or transform data represented as physical electronic or magnetic quantities within memories, registers, or other information storage, transmission, or display devices of a computing platform.

[0209] The system or systems described herein are not limited to any particular hardware architecture or configuration. A computing device may include any suitable arrangement of components that provides results conditioned on one or more inputs. Suitable computing devices range from general-purpose computing devices to specialized computing devices that implement one or more embodiments of the present subject matter, including general-purpose microprocessor-based computing systems that access stored software that programs or configures the computing system. Any suitable programming, scripting, or other type or combination of languages may be used to implement the teachings contained herein in software used to program or configure a computing device.

[0210] Embodiments of the methods disclosed herein may be performed in operation of such a computing device. The order of the blocks shown in the above examples may be changed, e.g., the blocks may be reordered, combined, and / or divided into sub-blocks. Certain blocks or processes may be performed in parallel.

[0211] Conditional language used herein, such as "can," "could," "might," "may," "for example," and the like, unless expressly stated otherwise or otherwise understood within the context in which it is used, is generally intended to convey that certain examples include certain features, elements, and / or steps, but not other examples. Thus, such conditional language is not generally intended to imply that features, elements, and / or steps are required for one or more examples, or that one or more examples necessarily include logic for determining, with or without author input or direction, that those features, elements, and / or steps are included in or performed in a particular example.

[0212] Terms such as "comprising," "including," and "having" are synonymous and are used inclusively in an open-ended manner and do not exclude additional elements, features, actions, operations, etc. Additionally, the term "or" is used in its inclusive (not exclusive) sense; for example, when used to associate a list of elements, the term "or" may refer to one, some, or all of the elements in the list. The use of "adapted" or "configured" herein means open and inclusive language that does not exclude devices adapted or configured to perform additional tasks or steps. Furthermore, the use of "based on" means open and inclusive in that a process, step, calculation, or other action that is "based on" one or more recited conditions or values may in fact be based on additional conditions or values beyond the recited values. Similarly, the use of "based at least in part on" means open and inclusive in that a process, step, calculation, or other action that is "based at least in part on" one or more recited conditions or values may in fact be based on additional conditions other than those recited or values beyond the recited values. Headings, lists, and numbering contained herein are for ease of explanation and are not limiting.

[0213] The various features and processes described above may be used independently of one another or may be combined in various ways. All possible combinations and subcombinations are intended to fall within the scope of the present disclosure. Furthermore, in some implementations, certain method or process blocks may be omitted. The methods and processes described herein are also not limited to a particular sequence, and the blocks or states associated therewith may be performed in other sequences as appropriate. For example, the described blocks or states may be performed in an order other than the order specifically disclosed, or multiple blocks or states may be combined into a single block or state. Example blocks or states may be performed sequentially, in parallel, or otherwise. Blocks or states may be added to or deleted from the disclosed examples. Similarly, the example systems and components described herein may be configured differently than described. For example, elements may be added, deleted, or rearranged compared to the disclosed examples.

Claims

1. 1. A method for determining tumor purity, comprising: obtaining nucleic acid sequence data representing a plurality of nucleic acid molecules of a biological sample of interest; aligning the nucleic acid sequence data to a reference genome; identifying a set of genomic regions based on the aligned nucleic acid sequence data, wherein each genomic region of the set of genomic regions comprises one or more nucleotide sequence variants relative to a corresponding genomic region of the reference genome; determining the B allele frequency of each genomic region of said set of genomic regions; determining a B allele frequency distribution for the biological sample based on the B allele frequencies for the set of genomic regions; processing the B allele frequency distribution using a trained machine learning model to estimate a metric that identifies tumor purity of the biological sample; and outputting the metric.

2. 10. The method of claim 1, wherein the nucleic acid sequence data is whole exome sequencing data.

3. The method of claim 1 , wherein the nucleic acid sequence data is whole genome sequencing data.

4. obtaining the tumor sample from the subject; and sequencing the plurality of nucleic acid molecules of the tumor sample to generate the nucleic acid sequence data; The method of claim 1 further comprising:

5. 5. The method of claim 4, further comprising isolating the plurality of nucleic acid molecules prior to sequencing.

6. identifying the set of genomic regions identifying one or more candidate nucleotide sequence variants in the nucleic acid sequence data; and 2. The method of claim 1, comprising calculating a reference read depth and an alternative read depth for each of the one or more candidate nucleotide sequence variants.

7. 10. The method of claim 1, wherein the trained machine learning model is trained with a training dataset generated from nucleic acid sequence data derived from one or more tumor cells diluted with normal cells.

8. The method of claim 1 , wherein the B allele frequency distribution is normalized.

9. 10. The method of claim 1, wherein the trained machine learning model has a mean absolute error of less than about 0.

2.

10. The method of claim 1 , further comprising outputting a report including information identifying the B allele frequency distribution.

11. 10. The method of claim 1, further comprising outputting a report including a metric identifying the estimated tumor purity.

12. 12. The method of claim 11, wherein the report further comprises information identifying at least one biomarker.

13. The method of claim 11 , wherein the report further comprises information identifying at least one prognostic marker.

14. 12. The method of claim 11 , wherein the report includes information identifying predicted somatic variants.

15. The method of claim 11 , wherein the report includes a treatment recommendation.

16. The method of claim 1 , wherein the tumor sample is from a human subject.

17. 16. The method of claim 15, wherein the treatment recommendation comprises a recommendation to not administer treatment to the subject.

18. 18. The method of claim 17, further comprising administering said treatment to said subject.

Images